acaps - ebola needs analysis project · acaps’ ebola needs analysis project (enap) would like to...
TRANSCRIPT
ACAPS - Ebola Needs Analysis Project April 2015
ACAPS - ENAP Liberia April 2015
2
Acknowledgements ACAPS’ Ebola Needs Analysis Project (ENAP) would like to acknowledge
all organisations and individuals who supported this assessment by
providing inputs to the questionnaires, assessment design and analysis of
the results. We would particularly like to thank InterNews, the Health
Cluster, Mercy Corps, the Ministry of Education (MoE), the National Civil
Society Council of Liberia, Save the Children, United Nations Children’s
Fund (UNICEF), United Nations Mission in Liberia (UNMIL) and Women’s
Campaign International for their support to the selection of appropriate key
informants (KIs).
Background to ACAPS, ENAP and Building Markets
The Assessment Capacities Project (ACAPS) is a non-profit initiative
backed by a consortium of three non-government organisations (NGOs),
Action Contre la Faim, Norwegian Refugee Council and Save the Children
International. It was created in December 2009, with the aim of supporting
the humanitarian community with needs assessments.
In November 2014, ACAPS launched a project to better analyse the
humanitarian impact of the Ebola epidemic in West Africa. ENAP provides
a comprehensive and independent picture of the impact of Ebola on the
needs of affected populations.
Building Markets (BM) was ACAPS’ implementing partner during this
assessment, and was responsible for identifying suitable KIs, data
collection and initial results processing. BM is a non-profit organisation,
which champions local entrepreneurs and connects them to new business
opportunities. Operating in six post-conflict and developing countries, BM
addresses the gaps between economic potential and access to
opportunities. BM has operated in Liberia since 2011 and administered
surveys during the Ebola outbreak on behalf of several donors, NGOs, the
UN, the World Food Programme (WFP) and US Agency for International
Development (USAID).
Contents
Acronyms ............................................................................................. 4
A. Executive Summary ........................................................................ 5
B. Background and Methodology ...................................................... 7
1. Outbreak evolution ......................................................................... 7
2. Background to the Assessment ..................................................... 7
3. Report Structure ............................................................................ 7
4. Methodology .................................................................................. 8
5. How to Read the Charts ............................................................... 11
C. Country level findings .................................................................. 13
1. Multi-sectoral prioritisation ........................................................... 13
2. Education ..................................................................................... 16
Context ........................................................................................... 16
Assessment findings ....................................................................... 16
3. Health .......................................................................................... 19
Context ........................................................................................... 19
Assessment findings ....................................................................... 19
4. Income Generation ...................................................................... 22
Context ........................................................................................... 22
Assessment findings ....................................................................... 22
D. County profiles ............................................................................. 25
1. Bomi ............................................................................................ 25
2. Bong ............................................................................................ 27
3. Gbarpolu ...................................................................................... 29
4. Grand Bassa ................................................................................ 31
5. Grand Cape Mount ...................................................................... 33
6. Grand Gedeh ............................................................................... 35
7. Grand Kru .................................................................................... 37
8. Lofa.............................................................................................. 39
9. Margibi ......................................................................................... 41
10. Maryland ................................................................................. 43
11. Montserrado ............................................................................ 45
12. Nimba ....................................................................................... 47
13. River Gee ................................................................................. 49
14. Rivercess ................................................................................. 51
15. Sinoe ........................................................................................ 53
Annex A – Questionnaire ..................................................................... 55
Annex B – Methodological annex ........................................................ 60
Acronyms
ACAPS Assessment Capacities Project
BM Building Markets
CBO Community Based Organisation
DFID Department for International Development
ENAP Ebola Needs Analysis Project
FAO Food and Agriculture Organisation
GoL Government of Liberia
DHS Demograhic Health Survey
IMS Incident Management System
INGO International Non-Governmental Organisation
IPC Infection Prevention and Control
KI (s) Key Informant(s)
KII (s) Key Informant Interview(s)
LISGIS Liberia Institute of Statistics and Geo-Information Services
MoE Ministry of Education
MSF Médecins Sans Frontières
NGO Non-Governmental Organisation
PI Personal Interview
SDP Supplier Development Programme (Humanity United)
SMI-L Sustainable Marketplace Initiative
UN United Nations
UNICEF United Nations Childrens Fund
UNMEER United Nations Mission for Ebola Emergency Response
UNMIL United Nations Mission in Liberia
USAID US Agency for International Development
WFP World Food Programme
WHO World Health Organisation
A. Executive Summary
Assessment background: At the beginning of April 2015, ACAPS
conducted a phone based, multi-sectoral assessment of KIs in all 15
counties of Liberia. The objective of the assessment was to identify the
main problems faced by communities, and how their needs differ from the
pre-Ebola situation. The assessment intends to inform the ongoing
discussions on early recovery and strategic decision making on sustainable
development.
Current priorities: Income generation, education and health are the main
problems faced by communities in Liberia, according to the 216 KIs
surveyed. The lack of job opportunities is the main obstacle for families to
obtain sufficient income. The subsequent inability to afford transport, school
materials and fees was identified by KIs as the main reason why children
are not attending school.
The impact of the Ebola outbreak on current priorities: The direct and
indirect consequences of the Ebola outbreak continue to impact the current
situation. The majority of KIs indicate that communities are currently worse
off when compared to the pre-crisis situation. There are two main
differences between current and pre-crisis needs: the continued impact of
the Ebola outbreak on livelihood opportunities and the fear of Ebola
infection at schools and health facilities.
The economy has started showing signs of recovery since Ebola incidence
decreased significantly, at the end of 2014. However, the assessment
results indicate that access to livelihoods is not yet at pre-crisis levels.
Business closures, reduced trade and general Ebola related
unemployment are still a top concern in most counties. KIs in Bomi, Bong
and Lofa stressed the reduction in agricultural production, the impact of
which will become clear during the harvest in October 2015.
Much of the response by development and humanitarian actors is currently
focused on infection prevention and control (IPC) in health and education
facilities. Response data indicates that 98% of targeted schools have
received IPC kits, for instance. Nevertheless, fear of Ebola transmission
remains a major concern in communities and prevents families from visiting
health facilities and sending their children to school. This fear has been
reported by almost all KIs countrywide, regardless of whether the county
has seen a high number of Ebola cases (see map next page).
At the same time, a number of KIs viewed the current situation as better,
when compared to the same time last year, particularly within the health
sector. The reasons provided for this improvement are the increased
support to health structures from the government and NGOs, and improved
awareness among communities on preventive behaviour and health
services. The increased NGO presence has generated more employment
opportunities, albeit temporarily. There is a high risk that positive impacts
of the Ebola response will be difficult to sustain as international support
recedes.
Suggested interventions: KIs suggest long-term interventions to address
the current problems. Training to improve the skills of families and increase
access to employment is the main suggested response, followed by
improvements to water, health, education infrastructure and roads. The
lack of payment of salaries and incentives for education and health staff
was highlighted as a main concern in half of the counties assessed.
Groups most in need of support: When asked to select three groups
most in need of support out of a list of ten, most responses concerned girls
under 18, Ebola orphans, followed by persons with a disability. In counties
with the highest number of reported Ebola cases, those affected by Ebola
were higher priority, with Lofa, Montserrado and Margibi reporting Ebola
survivors, their families and Ebola orphans as the main groups in need of
support.
ACAPS - ENAP Liberia April 2015
6
B. Background and Methodology
1. Outbreak Evolution
Liberia’s first two reported cases of Ebola were confirmed on 30 March
2014, in Foya district of Lofa County near the border with Guinea. By 7
April, the country reported 21 confirmed, probable, and suspected cases
and ten deaths. The situation in Liberia stabilised in April and through
May, with cases still largely concentrated in Lofa County. The
exponential growth of cases started when the first additional cases in
Monrovia were reported in mid-June. The city was not prepared to cope
with the outbreak and the disease spread rapidly.
On 6 August, President Johnson Sirleaf declared a three month state of
emergency. By 8 September, Liberia had the highest cumulative
number of reported cases in the region, reaching nearly 2,000 cases
and more than 1,000 deaths.
In late October, the first signs emerged that the situation in Monrovia
and Lofa had stabilised, with a slow decline in newly detected cases in
the early weeks of November. In mid-November, the government lifted
the state of emergency and set a target of no new Ebola cases by 25
December. During late November and early December, rural outbreaks
were reported as people who had been working in the cities returned to
their rural homes and by mid-December, the virus had largely moved
from cities to remote rural areas.
In February 2015, international borders, which were closed in August
2014, and schools, which were closed in July 2014, reopened. Since the
end of February only two new cases have been reported. If no
additional cases are confirmed, the country will be declared Ebola
free on 9 May, 42 days after the burial of the last confirmed case. (NYT
13/11/2014, WHO 01/2015)
2. Background to the Assessment
The current immediate priority for response actors in Liberia is to remain at
zero Ebola cases, by establishing and maintaining effective IPC measures.
Simultaneously, the national and international response is gearing towards
addressing the impact of the Ebola outbreak and underlying systemic
issues. To inform strategic decision making, ENAP conducted a
Department for International Development (DFiD) funded assessment in
Liberia. The survey aims to answer the following question: What are the
main problems currently faced by communities and how do they differ from
the situation before the Ebola outbreak? Comparisons are made primarily
between:
The current problems compared to the pre-crisis situation
The differences between and within counties
3. Report Structure The report presents the findings of the assessment and is divided into two
sections.
The first part (part C) covers the findings on a national level,
focusing specifically on multi-sectoral prioritisation, education,
healthcare and income generation. To provide these countrywide
priorities, the county specific responses were aggregated to a
national level. For more information on the methodology used,
please see Annex B.
The second part of the report (part D) outlines the county specific
results, highlighting if and how the situation within a specific county
is different from other counties.
ACAPS - ENAP Liberia April 2015
8
4. Methodology
The assessment process started mid-March and ended with the release
of the report on 29 April 2015.
Key informants (KIs)
The assessment was designed to collect data from primary sources at the
county level, through Key Informant interviews (KIIs). The number of KIs
sampled is proportionate to the population within a country:
A total of 218 KIs were interviewed, of which 216 were considered
sufficiently reliable (see data quality). There were some slight deviations
from the target number of KIs by county, as selection was to an extent
dependent on availability and eligibility. Efforts were made to capture
informants outside of the main urban areas, particularly in Montserrado,
where seven out of 32 informants were interviewed on the situation outside
of Greater Monrovia.
A large number of possible KIs were identified from existing traditional,
religious, governmental and NGO networks. Afterwards, potential KIs were
selected according to the following criteria:
Awareness of the situation of the entire population for their specific
county.
Knowledge of the needs that communities face, particularly in
accessing healthcare, education, and income generating
opportunities.
Understanding of how the current needs compare to the pre-Ebola
outbreak needs.
To test whether the respondents were sufficiently knowledgeable to speak
for a specific county, several eligibility questions were asked before the
start of the actual survey:
What is your position within your county and what is your
engagement with communities in your county?
Timeframe
Assessment preparation 16–25 March
Training 26–27 March
Data collection 31 March–8 April
Data processing, analysis and report writing 9–29 April
Number of key informants
Population size No of KI targeted
<50,000 3
50,00 to 100,000 10
100,000 to 500,000 15
>500,000 30
ACAPS - ENAP Liberia April 2015
9
What is the name of the senator in your county?
What are the main sources of income within your county?
In your opinion, how knowledgeable are you about the situation of
communities in your county? Do you know about the situation in
the whole county?
Out of the 330 informants approached, 20 were
not considered eligible, while over 80 could not
be reached as their phone was either switched
off or the number provided was incorrect.
The average age of respondents was 43 years.
32 informants were under 30 years of age. 25%
of the KIs were female. 60% of informants had a
university degree, while 36% finished secondary school but did not finish
university.
To ensure multiple perspectives were captured, KIs from at least three
different institutions were targeted by county. This included government
officials, staff working for
community based
organisations (CBOs)
and NGOs, media and
traditional leaders:
1 The first cases in Liberia were confirmed at the end of March 2014. However, at that time the outbreak was concentrated in one part of the country (Lofa County). The crisis escalated when the outbreak spread to Monrovia in mid-June.
Questionnaire
A multi-sectoral questionnaire was designed to capture:
Current priority issues and needs for intervention
The impact of the Ebola outbreak on the current situation
How the situation differs within and between counties
Specific vulnerable groups
The questionnaire focused on access to education, health and livelihoods.
These sectors were selected because:
It is generally assumed that the Ebola outbreak still has an impact
on these needs
There is limited information available on these sectors and, at the
time of the survey preparation, no comprehensive sector-specific
assessment was planned.
Sensitive questions, including those related to protection and symptoms of
Ebola, are unlikely to yield useful data during phone based interviews and
were therefore not included
KIs were requested to compare the current situation to that existing before
the Ebola crisis escalated1, taking the end of March 2014 as the baseline
reference point to ensure that seasonal variations were controlled.
Data quality
Although phone based surveys have several distinct advantages, for
instance allowing a large number of households (HHs) to be surveyed in a
relatively short period of time, there are also particular constraints. One of
the main constraints is that it is challenging to judge the reliability of the
answers provided by respondents, as findings cannot be corroborated by
direct observation. Several steps were taken to improve data confidence:
ACAPS - ENAP Liberia April 2015
10
Quality and diversity of the informant: KIs were selected according
to their knowledge of the situation, following three criteria. During
the interview, their eligibility was tested (see key informant profile).
An effort was made to ensure a diverse sample, in terms of
respondent background, geographic coverage, gender and age.
After every questionnaire, the researcher was asked to rank their
confidence in the accuracy of the responses, on a scale of 1 to 3.
The responses of two KIs were discarded as they were deemed not
reliable.
Confidence level
1 Answers provided appear consistent / no reason to doubt accuracy
of responses
2 Respondent might not have understood all questions / some reason
to doubt responses
3 Respondent clearly did not understand most questions / significant
reason to doubt responses provided
Triangulation: Two types of triangulation of information provided by
informants took place. For every county at least ten KIs were
interviewed, allowing for corroboration of findings. In addition, the
information obtained was complemented and triangulated with
secondary sources. These included baseline reports and datasets
from the Government of Liberia (GoL), UN Mission for Ebola
Emergency Response (UNMEER), Médecins Sans Frontières
(MSF), World Health Organization (WHO), UN Development
Programme (UNDP), WFP and other humanitarian and
development organisations.
Analysis
Four types of analysis were applied to the data:
Aggregation of responses to the national level: Respondents
solely spoke about the conditions in their counties of residence. To
obtain a nationwide perspective, the county level response was
weighted to reflect the population size of the area KIs reported on.
The weighting is calculated by dividing the total county population
by the number of KIs interviewed about the county. For more
information on the calculation of national averages, please see
Annex B.
Differences between and within counties: KIs in all 15 counties
were requested to provide their perspective on the situation within
their county, thereby allowing for a comparison of needs and
priorities between the different counties. In addition, KIs identified
districts within the counties where communities faced above
average needs. The top three districts mentioned (and more if tied)
were selected as priority districts.
Access vs Availability: The main problems mentioned by KIs were
categorised as ‘access’ and ‘availability’ related obstacles. Access
to basic services is defined as both physical and economic access,
for instance to education, healthcare and income generating
opportunities. Availability refers to whether the services are present
and of sufficient quality. The categorisation supports response
planning as it provides guidance on whether interventions should
focus on improving the existing infrastructure, provide support to
families to access the infrastructure or a combination of both.
Before and After: The questionnaire was designed to capture KI
views on the current situation, and how these problems compare to
pre-outbreak needs. This analysis provides an indication of the
direct and indirect consequences of the Ebola outbreak and to what
extent this impact remains a relevant factor to consider while
planning interventions.
ACAPS - ENAP Liberia April 2015
11
Limitations of the data and survey methodology
Although efforts were made to ensure diversity of the sample, it is
likely that the situations of some segments of the population have
not been captured. For instance, it is likely that KIs were not aware
of the situation of all those communities in remote and hard to
access areas.
The majority KIs interviewed (75%) were male. This gender
imbalance should be kept in mind when interpreting the results. A
preliminary gender specific analysis indicates that there are slight
differences between the perspectives of male and female KIs. For
example, 25% of female respondents indicated that girls under 18
are in particular need of support, compared to 20% of male KIs.
KIs were selected on the basis of their knowledge of the wider
community and situation. As a result, they enjoyed a higher level of
education than average, and did not represent a true cross-section
of the population.
Although the findings can be considered indicative of the conditions
in a county, interpreting small differences in local priorities as a
basis for different recovery policies should be avoided as policy
should focus on robust differences.
5. How to Read the Charts
This section offers some tips as to how to read and interpret the main charts
used in an appropriate way, to fully understand the findings. In addition to
maps, three types of charts are used to visualise the findings:
Heat-maps to summarise priority problems
Proportional text based visuals to reflect before and after
conditions
Horizontal stacked bar charts to summarise sector specific
obstacles, suggested interventions and vulnerable groups.
“Priority” or “preferences” visuals
Heat-maps are used to summarise multiple priority responses and their
relative importance, into a form that is easier to visualise. The questions
from which the heat maps are extracted always imply a selection of the top
problems.
EXAMPLE
Levels of preference are grouped under several sub-headings. They are
sorted in descending importance order on a national level and in
alphabetical order on a county level, to allow for easy comparison between
counties. The darker the colour, the more frequently the response was
mentioned.
In the example heat-map, results can be interpreted as follows: Income
generation was ranked as the main problem, followed by education and
health. Water and access to food were ranked lowest.
Before and after conditions
Text based visuals are used to reflect the answers to questions such as
“How would you describe the means for communities to generate an
income before the crisis and how would you describe the current
conditions, compared to the same time last year?” The size of the text is
proportionate to the percentage of KIs providing the specific response.
ACAPS - ENAP Liberia April 2015
12
EXAMPLE
In the above visual, results can be interpreted as follows: when weighed
for population size, the predominant view of KIs is that the livelihood
situation has become worse when compared to the pre-crisis situation.
Prioritisation of districts
The county level maps illustrate the answer to the question “In which
districts in your county do people currently face the most problems?” Only
those districts that we’re mentioned by a significant number of KIs were
considered as a priority. To calculate those priority districts, the top three
(or more if tied) identified districts were selected. The proportion of KI
responses for these districts was divided by 2.5 to limit the possible bias of
individual respondent’s preferences, and if the resulting number fell above
a threshold value the district was marked as a priority.
EXAMPLE
In Grand Gedeh, for instance, the districts of Gboe-Ploe, Tchien and
Konobo were identified by KIs in the county as areas where communities
are facing the most problems. The population figure of 144,872 refers to
the people residing in the county, according to 2015 government
projections.
C. Country Level Findings
1. Multi-Sectoral Prioritisation
KIs were asked to rank the problems in water, health services, education,
livelihoods, food and safety and security in priority order across their
county. Results were then aggregated to give country-wide priorities.
Overall, income generation was ranked as the largest current problem. The
endemic lack of job opportunities is exacerbated by the economic impact
of the Ebola outbreak on employment levels. The reduction of livelihood
opportunities is the most significant change from the pre-crisis outbreak,
perceived by KIs. Education and health were also ranked as a high priority,
both before the crisis and currently.
Food was ranked as the lowest priority across the board. However, with
many families engaged in subsistence farming, access to food and income
generation are closely linked in Liberia. It could be that the structural levels
of food insecurity in Liberia were perceived by some KIs as an income
generation concern, and not mentioned separately. For more information
on the access to food as an ‘irrelevant alternative’, see Annex B.
In most counties the prioritisation follows the national results, but in two
counties the perceptions of KIs on the main priorities differ significantly. In
Bong, access to water was ranked high, which is likely to be linked to the
very low quality of the water infrastructure before the crisis. In Lofa, access
to food was prioritised as a main current problem. Lofa is one of the few
self-sufficient counties in terms of food production. As a result, any
decrease in agricultural production has a direct impact on food availability,
with limited opportunities for families to access alternative sources. This
assessment did not cover the food security situation in-depth and further
research on the food security situation in Lofa is required.
Suggested interventions
One of the objectives of the survey was to capture KI perspectives on the
required interventions. The intervention most often suggested is training to
improve the skills of families, to address the current problems. According
to KIs, an increase in skills and literacy among families will improve access
to income generating opportunities. The resulting increase in income will
also lessen existing financial obstacles to education and healthcare.
ACAPS - ENAP Liberia April 2015
14
These suggestions do not fully address the identified problems. Although
skill development will increase access to employment, a continued lack of
jobs will to prevent families from significantly expanding their income
sources. The obstacles mentioned indicate that KIs consider access to
basic services, such as education and healthcare, as more pronounced
than availability. However, the suggested interventions primarily focus on
the need to improve health, education and road infrastructure, which
indicate the importance of interventions which address both dimensions of
the existing problems. The discrepancies between the main obstacles
mentioned and main interventions proposed illustrate that, although it is
important to integrate KI priorities into response planning, these views
should be complemented with expert judgement on the most appropriate
health, education and livelihoods interventions.
Vulnerable groups When asked to select three groups most in need of support out of a list of
ten, most responses concerned girls under 18. Secondary data highlights
the specific vulnerabilities of girls under 18: for instance, an assessment in
2013 found that 25% of women aged 15–18 were either pregnant or had
given birth (DHS, 2013). The identification of girls under 18 as a vulnerable
group was even more pronounced among female KIs, with 25% of female
respondents highlighting this group as in particular need, compared to 19%
of male KIs.
Ebola orphans were mentioned as the second priority group most in need.
Over 3,000 children have been registered as having lost one or both
parents due to the Ebola outbreak. As registration is currently still on-going,
the actual number is likely to be higher (UNICEF, 15/04/2015; Child Protection sub-
Cluster, 27/04/205 (unpublished)).
Persons with a disability were mentioned as one of the main groups in need
of support as well. During a 2010 assessment, 4% of the Liberian
population over 5 years of age reported at least one disability, caused
primarily by accidents, the civil wars and birth defects (Labour Force Survey, 2010).
There are noteworthy differences between counties. In counties with the
highest Ebola incidence, those affected by Ebola were prioritised higher
than in other countries. KIs in Lofa, Montserrado and Margibi reported
Ebola survivors, their families and Ebola orphans as the main groups in
need of support. Surprisingly, Rivercess and Sinoe also mentioned these
groups within their top three, although Ebola incidence has remained low
in these counties. Pregnant women are specifically highlighted as one of
the main groups in need in Maryland, River Gee and Grand Gedeh.
2. Education
Context
Pre-crisis situation
Primary and secondary education is compulsory from the ages of 6–13
(UNICEF, 09/2012). 2013 data indicates that 38% of children aged 6–11 were
attending primary school, ranging from 54% in Montserrado to 19% in
Bong. 23% percent of youths aged 12–17 attended secondary school. The
actual attendance numbers at primary school are much higher, as a large
number of children outside of the official school age range are attending
school. There is a stark difference between attendance rates in
Montserrado and other counties, with girls and boys in urban areas much
more likely to achieve higher levels of education (DHS, 2013). Basic education
(between 6–11 years’ of age) is free in all government schools. However,
this policy has not been adopted by all schools in the country and there are
additional costs for uniforms and school materials. There is a registration
fee for secondary school. School fees / costs were consistently identified
as the primary barrier to education pre-Ebola, followed by the need for
children to stay home to support the family. (Liberia Education Cluster, 17/03/2015;
Government of Liberia, 2015)
Ebola impact
In July 2014 the GoL decided to close all schools to prevent Ebola
transmission. As a result, an estimated 1.4–1.5 million school aged children
were out of school between July 2014 and February 2015. An assessment
by the Education Cluster showed that families consider contracting Ebola
and the costs of education as the main obstacles to returning to schools,
once they reopen. In the same assessment, students highlighted corporal
punishment as a main concern. School administrators also identify traffic
on the way to and from school as a major risk. Once the schools opened at
the start of March, many parents reportedly refused to send their children
to school, due to the fear of Ebola transmission (Liberia Education Cluster,
17/03/2015; USAID, 28/01/2015; UNICEF, 04/03/2015; UNMEER, 02/03/2015). The
availability of WASH facilities, a key component of Ebola prevention,
remains low in schools. 31% of schools assessed by the Education Cluster
did not have functioning latrines, while schools that do average one latrine
per 123 students. Only 60% of the schools had safe drinking water within
500m, 40% of schools had soap and water for hand washing, and 39% had
functional hand-washing facilities. During the long period of school
closures, many of the schools were looted or damaged. 74% of the 351
schools assessed by the Education Cluster reported having had some type
of school material lost, stolen or damaged since the beginning of the Ebola
crisis (Education Cluster, 17/03/2015).
Assessment findings
Most children attended primary school before the Ebola outbreak,
according to the 216 KIs. However, other data sources indicate that less
than half of school aged children attended primary school nationwide.
Throughout the country, the proportion of children going to school is
overestimated by KIs, when compared to other sources. This could indicate
a misunderstanding of the term ‘most’ in this context. The majority of KIs
indicate that the situation is worse compared to a year ago, and this view
is supported by multiple other assessments. Those that considered the
situation to have improved assign this to a stricter enforcement of the free
education policy and NGO support to teachers and education facilities.
Priority obstacles
The lack of access to education was ranked as one of the main problems.
The main obstacle to education was the inability of families to pay for
transport, fees and education materials, before the crisis and currently. An
additional current obstacle is the
perceived risk of Ebola infection
in schools, which discourages
families from sending their
children. As can be seen on the
map on page 6, the level of fear
is not directly linked to Ebola
incidence in the counties. It can
be explained by a lack of IPC at
facility level, active social
mobilisation spreading fear of
Ebola, or misinformation and
rumours about the causes of
Ebola.
Overall, the low quality of school
infrastructure is currently ranked
as a larger concern compared to
the pre-crisis situation. This is likely due to the damaging and looting of
school infrastructure during the extended period of closure. The low quality
of infrastructure was specifically highlighted by KIs in Grand Kru.
In three counties, Gbarpulo, Lofa and River Gee, KIs highlighted that
students and families did not see the added value of education before the
outbreak of the crisis. Only in River Gee does this continue to be one of the
main obstacles. The change over time is either influenced by social
mobilisation, which accompanied the back to school campaign, or the
relative higher importance of other existing problems.
The large majority of schools have reopened, according to a school
monitoring exercise by the Education Cluster. Only in Gbarpolu does the
closure of schools or unavailability of staff remain a significant obstacle,
according to KIs.
Access vs Availability
To better understand the kind of problems communities face, the obstacles
were categorised into ‘availability’ concerns (whether adequate education
infrastructure, including staff and furniture, was available) and ‘access’
concerns (whether children could access the required transport, financial
means and information to be able to attend school). Problems that can be
categorised as a lack of access to education are currently the main
obstacles to children attending school. The main pre-crisis access
considerations remain. The fear of Ebola infection has become a more
prominent concern for families not sending their children to school, than the
need for children to support their families The availability of education is
primarily an issue in Bomi, Rivercess, Sinoe and Grand Kru, although for
very different reasons. In the first three counties the low quality of school
infrastructure was highlighted as a key obstacle, while in Grand Kru the low
number and quality of staff is a major problem. Illustratively, the training of
teachers is highlighted by KIs in Grand Kru as the most important
intervention currently required.
Interventions
Improvement of school infrastructure and regular payment of school staff
are the main interventions proposed by KIs to address the existing
problems. They particularly stressed the need for the payment of salaries
in Bomi, Bong and Grand Gedeh.
ACAPS - ENAP Liberia April 2015
19
3. Health
Context
Pre-crisis situation
The Ebola outbreak exposed the structural weaknesses of the healthcare
system in Liberia. With 8.6 skilled health professionals per 10,000 in 2010,
Liberia is among the lowest in the world2 (Liberia Health System Assessment, 2015).
Delays and inefficiencies in the payment of salaries, allowances and
incentives have persisted over the years, resulting in further critical
shortages of qualified staff. Illustratively, in February 2014, before the
outbreak was declared, health centre staff went on strike because of lack
of payment. At 6.2% of GDP in 2013–14, government expenditure on health
falls below the agreed 15%3. The suspension of user fees for health
services in 2006 has only been moderately successful (Liberia Health System
Assessment, 2015). Transport from remote areas to health centres is often
expensive and lengthy due to the poor road network. The lack of
infrastructure also affects the supply chain, resulting in a lack of medicine
and medical equipment. Shortages of health supplies tend to be highest
in counties in the southern and eastern parts of the country (Liberia Health
System Assessment, 2015 (unpublished)). The 2013 Demographic Health Survey
(DHS) showed that over 60% of women face at least one problem
accessing healthcare, mostly a lack of money or long distances to health
facilities (DHS, 2013). The lack of access to quality healthcare resulted in high
levels of morbidity and mortality. In 2013, maternal mortality rates were as
high as 1,000 maternal deaths per 100,000 live births, making it one of the
highest in the world (HRH, 12/12/2014; IMS, 30/01/2015; UNICEF, 29/10/2014).
Ebola impact
2 WHO recommends a minimum density threshold of 22.8 skilled health professionals/10,000 people to provide the most basic health coverage. WHO, 01/2014
During the height of the outbreak, a large number of health facilities closed
or decreased their service provision. At one point in August for instance, all
five of the main hospitals in Monrovia were closed (MSF, 15/08/2014). While
most have reopened since the decrease in Ebola cases, most facilities
were providing reduced services at the start of 2015, compared to before
the Ebola outbreak (PI, 19/02/2015; UNMEER, 29/12/2014). The disruption of services
was primarily caused by the lack of staff. At least 370 health workers were
infected (probable and confirmed cases) and over 180 died, as at April
2014. Many other health workers stopped working out of fear of infection.
In addition, routine health care workers were not paid between September
2014 and March 2015. The high number of health care workers infected
has eroded public trust and communities became afraid of seeking
treatment (IMS, 02/04/2015 (unpublished); IMS, 23/01/2015).
Assessment findings
To obtain an overview of how the Ebola outbreak has affected access to
health services, the 216 informants were asked to judge the health situation
before the Ebola outbreak and how the situation has changed since. Before
the crisis, in most areas, the majority of people went to a health facility for
treatment. For half of the population assessed, the current health situation
was worse compared to last year. However, for a significant number of KIs
the situation has improved, including all KIs in Bong County. This is
primarily because of the additional support provided to the health structure
as a result of the outbreak. KIs mentioned increased awareness raising,
the training of health staff and provision of supplies as particular activities
that have led to an improvement.
3 Abuja Declaration, 2001 (WHO, 2011)
Priority obstacles
When asked about the top three problems across all sectors, KIs indicated
that access to healthcare was the one of the main problems before the
crisis. Currently, health has become less of a priority as compared to other
sectors, after income generation and access to education.
There are large variations
between counties. KIs in
Sinoe, Lofa and Bong did not
consider health a major
concern while those in Grand
Kru ranked it as the main
issue.
Before the outbreak, the lack
of medicine was ranked as
the main obstacle to obtaining
healthcare, throughout the
country, followed by the lack
of financial resources to pay
for health services.
Currently, these problems remain the main obstacles, with a lack of
medicine, money and logistical constraints ranked in the same order as
before the crisis. However, the fear of Ebola infection in health facilities has
become by far the largest concern, with two thirds of KIs mentioning this as
an obstacle.
Traditional health care was the only reliable source of care during the civil
wars, between 1989 and 2003. Recent improvements to the formal health
care system have reduced reliance on this sector, but it remains an
important source of care, particularly during birth (Kruk. Et al, 2011). However,
the responses indicate that families are less likely to choose traditional
healers currently, compared to last year. This could be explained by the
fact that several localised outbreaks were sourced to traditional healers,
which could have eroded trust in their practices. In addition, part of the
social mobilisation messaging during the outbreak warned patients with
Ebola symptoms against the use of traditional healers.
Access vs Availability
To understand whether health interventions should focus on improving the
health infrastructure or enabling families to visit health centres, the findings
were categorised as ‘availability’ obstacles (e.g. the lack of health centres)
and ‘access’ obstacles (e.g. families do not have the means to access
available health centres). Both before the crisis and currently, obstacles to
health care are mostly driven by access issues, for instance a lack of money
and fear of Ebola hampering families from visiting health facilities.
Meanwhile, the availability of health services remains insufficient. The
assessment results show that pre-crisis problems related to the lack of
qualified staff have been compounded during the crisis. Non-payment of
salaries and a decrease in the number of staff, due to Ebola infections, are
often mentioned as problems.
Interventions
KIs were asked to identify the three main activities required to address
problems in their county. Most health related interventions proposed were
intended to strengthen and expand the health infrastructure. The need to
improve the quality of the health infrastructure and provide supplies, such
as medicines, were the most often mentioned interventions. KIs in River
Gee specifically mentioned the need for additional health supplies, with half
of KIs mentioning this as one of the main activities required. The need to
pay health staff salaries was highlighted by KIs in nine of the 15 counties,
most often in Bomi and Bong.
4. Income Generation
Context
Pre-crisis situation
The main source of income varies significantly between urban and rural
Liberia. The agricultural sector is the primary livelihood source in rural
areas, mainly smallholder and subsistence farming and the production of
cash crops such as rubber, palm oil, cocoa and coffee. Rice is the main
staple food grown, followed by cassava (FAO, 2012). Agricultural productivity
is low due to a lack of modern technology, inadequate tools, limited access
to markets, and a lack of access to capital / credit.
Most urban HHs derive their income from two or more livelihood sources,
combining agriculture with petty trade, hunting and casual labour (WFP, 2013;
OECD, 2008; FAO/WFP, 05/01/2015; DHS, 2013). In Greater Monrovia, sales and
services are the most dominant livelihood group, employing over 80% of
working age men in 2013 (DHS, 2013). The informal sector, which includes
petty trading, construction, mechanics, and security services, remains one
of the major sources of employment and income for the urban population
in Greater Monrovia (WFP/GoL, 2010).
Unemployment estimates vary greatly. The most recent data, from 2013,
indicates that around 23% of working age men and 44% of working age
women were unemployed in the 12 months preceding the survey (DHS, 2013).
Under-employment is extremely high, at over 60% (WFP, 10/2010). The large
majority of Liberia’s youth missed out on basic education during the civil
wars, therefore today’s adults lack the required skills to obtain gainful
employment in either the private or public sectors (WFP, 06/2013).
Ebola impact
The largest economic effects of the outbreak have been those linked to
changes in behaviour, either as part of government-imposed aversion
measures or driven by fear of Ebola infection.
The GoL imposed a state of emergency between 6 August and 13
November 2014. Borders and markets were closed, transport restrictions
and curfews were imposed on some communities and counties, and
government activities were limited. At the same time, many NGOs and
large mining and agricultural companies scaled down their operations and
evacuated their expatriate staff for fear of contracting Ebola (ACF/WHH, 12/2014;
FAO/WFP, 12/2014; BM, 03//2015). According to a survey in October and November
2014, roughly half of urban respondents and a third of rural respondents
were no longer employed, despite working earlier in the year. This survey
also found no statistically significant differences in unemployment levels
between the worst-affected by Ebola regions and other regions of the
country (FEWSNet, 11/2014). At the start of 2015, trading activities and the
general economy started showing signs of recovery, with several
assessments indicating that there has been a substantial return to work. At
the end of February, the Liberian side of the international borders reopened (FAO/WFP, 12/2014; BM, 03/2015; WB, 24/02/2015).
Assessment findings
Despite the initial signs of recovery, two thirds of KIs still consider the
current conditions worse than they were pre-outbreak. KIs in five counties
hold a different perspective - most KIs in Bomi, Bong and Grand Gedeh
indicate that the current situation no longer differs from the pre-crisis
situation. Several KIs judged the situation to have improved, compared to
the year before. In Maryand, Grand Kru and Grand Gedeh, they point to
the increased availability of employment due to the presence of NGOs as
a cause for the improvement.
ACAPS - ENAP Liberia April 2015
23
Priority obstacles
Although there have been some signs of market recovery since Ebola
transmission has slowed down, the assessment findings show that the
labour market remains heavily affected by the secondary impacts. KIs
nationwide rank access to income generation as a larger problem than
access to healthcare, education, water, safety and security and food.
The lack of jobs, one of the main concerns before the outbreak, remains
one of the main obstacles to obtaining an adequate standard of living. This
problem has further
expanded due to the
outbreak, with the
decrease in agricultural
production and closure or
disruption of markets and
businesses.
The only county where the
endemic lack of job
opportunities is not the
main current concern is
Lofa. KIs in Lofa, which
was the epicentre of the
early Ebola outbreak,
indicated that increased
unemployment due to the
outbreak and the impact
on markets and
businesses is the main
current obstacle.
The Ebola outbreak peaked during a time when famers would normally
plant rice. Movement restrictions are likely to have led to a decrease in
yields. The decrease in agricultural production was mentioned by KIs as
one of the top three obstacles in Bomi, Bong and Lofa.
Stigmatisation of Ebola survivors was mentioned as a concern in
Montserrado and Lofa, the counties most affected by the outbreak.
Access vs availability
Existing availability problems, primarily the lack of jobs, have become more
prominent than access related concerns due to the continued negative
impact of the outbreak on economic productivity. Access to income
opportunities is currently mainly influenced by the lack of literacy or skills.
This is a structural problem, primarily affecting youth who had limited
access to education during the civil wars.
Interventions
Training to improve skills is the intervention most required to address
existing problems, according to KIs. The need for additional support to
agricultural production was mentioned in River Gee and Lofa, while the
need to improve the road infrastructure, which will facilitate trade, was the
third most suggested intervention. Overall, KIs offered limited suggestions
on how to address the lack of income generating opportunities, although
the lack of job opportunities is the main factor hampering access to income
and, by extension, access to education and health care.
ACAPS - ENAP Liberia April 2015
25
D. County Profiles
1. Bomi
Key findings
The current situation in Bomi does not differ significantly from before the
crisis, according to KIs. Income generation, education and health were the
main priorities before the outbreak, and are now. The lack of salary
payments to teachers and health staff seems to be a more pressing issue
in Bomi than in other areas of Liberia, as does the perceived reduction in
agricultural production. KIs highlighted three districts as being most in need
of support, Senjeh, Dowein and Suehn Mecca. These districts are
traditionally less accessible (GoL, 2008) and the limited number of roads,
health facilities and schools are the main reasons why these districts were
selected over others in the county.
ACAPS - ENAP Liberia April 2015
26
Context
Bomi is the most food insecure county in Liberia, despite being rich in
natural resources and having a large agricultural potential. A 2013
assessment found that 55% of the population were food insecure,
compared to 18% nationwide, the highest proportion in the country (WFP,
2013). The remoteness of villages, lack of employment opportunities and
limited food production contribute to the high levels of food insecurity.
Illustratively, only 33% of the male population was engaged in agriculture
in 2013, compared to 62% of the population in all districts in rural Liberia.
In 2008, 105 villages, or an estimated 20,000 people, were reportedly cut
off from vehicular transport (GoL, 2008). In 2013, school attendance rates in
Bomi were the fourth lowest in the country, at 31% of children aged 6–11.
The first Ebola case was reported from Bomi mid–May, making it the
second county affected by Ebola after Lofa. Over 130 confirmed Ebola
cases were reported between May and the end November 2014 (MoH, 04/2015
(unpublished)). The local infrastructure was overwhelmed by the outbreak in
September 2014, the two largest hospitals were temporarily closed, leaving
over 330,000 people without healthcare (NYT, 12/09/2014).
Education
As in the rest of the country, the lack of resources is the main reason why
families were not able to send their children to school before the crisis. This
is still the main concern, followed by the fear of Ebola infection in schools.
The low quality of teachers or school staff was of larger concern in Bomi
than other counties, with 40% of KIs mentioning this as one of the three
main obstacles, compared to 10% countrywide. To address this problem,
KIs proposed regular payment of teacher salaries.
Healthcare
With 2.5 health facilities per 10,000 residents, Bomi County is relatively well
served in terms of health infrastructure compared to other counties (Liberia
Health System Assessment, 2015). The majority of KIs agree that most people
accessed health facilities to seek treatment before the crisis. The
availability of staff was mentioned as one of the main concerns, both before
the crisis and currently. Ten out of 15 KIs highlighted the need to pay health
staff salaries. 80% of KIs indicated that the current situation is similar or
better than the same time last year. Increased assistance to health facilities
following the Ebola outbreak and training of healthcare workers are the
main reasons mentioned for this improvement. As in the rest of the country,
the fear of Ebola infection is the main obstacle to accessing healthcare.
Income generation
Obtaining sufficient income to support families was a significant concern
before the crisis. KIs continue to rank the lack of job opportunities as the
main obstacle to generating an adequate income. The closure / disruption
of markets is another major obstacle to finding employment. One KI
mentioned that the presence of NGOs responding to Ebola has lowered the
rate of unemployment in the county. The decrease in agricultural production
due to Ebola is ranked relatively high when compared to the rest of the
country, and currently forms one of the most important obstacles to
generating an income. As one of the areas with a high number of Ebola
cases, it is likely that farmers in Bomi reduced the number of workers used
for planting and harvesting. Additional data collection is however required,
to see to what extent agricultural yields have decreased.
ACAPS - ENAP Liberia April 2015
27
2. Bong
Key findings The views of KIs in Bong differ significantly from those in the rest of the
country. KIs consider the current situation similar or better than a year ago,
with all informants highlighting that the health situation has improved. This
is primarily attributed to increased NGO and government support, as a
result of the Ebola outbreak.
The main priority is access to water, according to KIs, both currently and
before the crisis. This is a much lower priority in other counties. Salary
payments for health staff and teachers are a larger concern to KIs in Bong,
compared to the rest of the country. The fear of Ebola, a main concern in
most of the counties of Liberia, was not ranked as a priority.
ACAPS - ENAP Liberia April 2015
28
Context
Bong County has one of the largest proportions of people working in
agriculture, with 70% of men engaged in cultivation in 2013, compared to
40% nationwide. The civil wars (1989–2003) destroyed much of the
infrastructure and economy, resulting in high levels of unemployment,
especially amongst young women and men (GoL 2008). The reconstruction
and development of iron ore mines was delayed, partly due to the global
economic crisis in 2008–09. The county was one of the few in the country
where poverty actually increased, to 67% between 2007–10 (WB, 2011).
In 2013, Bong had the lowest primary school attendance rates in the
country, 19% of children aged 6–11. Most of the school buildings in the
county were damaged or destroyed during the civil war. Literacy rates were
the lowest in the country in 2013, at 53% compared to 70% countrywide (DHS, 2013; GoL, 2008).
At one facility per 10,000, the county experiences the lowest health facility
coverage in the country (DHS, 2013; GoL, 2015). However, the main hospital,
Phebe hospital, was known as one of the best hospitals in rural Liberia and
receives significant national and international support (CNN, 24/09/2014; PI,
24/04/2015).
Water infrastructure is traditionally underdeveloped in Bong, and with one
of the highest rates of open defecation in the country (73% according to the
2013 DHS), the quality of water is of concern (DHS, 2013; PI, 24/04/2015).
The first Ebola case was reported in July 2014, and 149 additional Ebola
cases were reported in the following five months (MoH, 04/2015).
Education
Over half of KIs judged the current education situation to be better than a
year ago, primarily due to the reduction and abolishment of fees and
increased government support to education facilities. However, access to
education before the crisis was one of the worst in the country and
significant concerns remain. The main obstacles to accessing education, in
order of priority, are the lack of resources for families, the low quality of the
infrastructure and logistical constraints, according to KIs.
Health
All KIs agreed that the current health situation is better compared to the
year before. Additional health facilities have been established, more health
care workers are working in the county and there is increased awareness
among communities on the importance of health. The lack of medicine, a
main concern before the crisis, seems to have been partly addressed by
the health interventions. KIs no longer mention this as a problem. The lack
of payment of staff salaries and incentives, a concern before the outbreak,
is currently ranked as the main obstacle to accessing treatment by KIs.
Multiple KIs warned that health care workers will go on strike if the
outstanding payment issues are not resolved.
Income generation
The main income generation issue is the lack of job opportunities, pre-
Ebola and currently. Bong is one of the three counties, in addition to Bomi
and Lofa, where the decrease in agricultural production was prioritised as
a significant problem by KIs. Secondary data indicates that the closure of
markets and reduction in trade was a major concern at the height of the
outbreak (WFP/FAO 12/2014). However, KIs indicated that since the reopening
of the roads and the markets, the situation has improved significantly.
ACAPS - ENAP Liberia April 2015
29
3. Gbarpolu
Key findings
The overall situation in Gbarpolu is currently not significantly different from
the pre-crisis conditions, according to the prioritisation of problems by KIs.
Access to education and health ranked as the main concerns before the
crisis and at the time of the assessment. However, the secondary impacts
of the Ebola outbreak, including the fear of Ebola, continue to affect the
needs in the county. The main interventions suggested by KIs focused on
the structural causes of the current problems, including the improvement of
health, education and road infrastructure. KIs highlighted Belleh,
Gounwolaila and Kongba as areas where communities face more obstacles
than in other areas, mainly due to the lack of job opportunities and health
infrastructure.
ACAPS - ENAP Liberia April 2015
30
Context
Like the rest of rural Liberia, the main source of income in Gbarpolu is
agriculture, followed by unskilled manual labour. Gold and diamonds are
the most commonly exploited mineral resources in the county.
The county is specifically underdeveloped in terms of access to education
and had the lowest literacy rate of the country in 2013 (GoL, 2008; DHS, 2013). In
2013, 40% of the population was part of the poorest 20% in the country.
However, food insecurity levels in the county match the national average,
which was 20% before the crisis (WFP, 06/2013). Most families cultivate food
crops or engage in fishing (GoL, 2008).
Gbarpolu was one of the last counties where an Ebola case was confirmed,
six months after the first case in Liberia. 15 additional cases were
confirmed, with the last being reported at the end of September 2014 (MoH,
04/2015).
Education
Education has been ranked as one of the main concerns in Gbarpolu, both
before the crisis and currently. In 2013, school attendance rates were
among the lowest in the country (26% of children aged 6–11) (DHS, 2013).
Over half of KIs (60%) agree that the situation has become worse.
The fear of Ebola infection in schools is the main obstacle to accessing
education, despite it being over six months since the last confirmed case.
The results suggest that the reopening of schools has not been as effective
in Gbarpolu, compared to the rest of the country. 30% of KIs in Gbarpolu
mentioned the closure of schools or staff not returning to their work as an
issue, compared to 10% across the country.
Health
Health is one of the other main problems, according to KIs. As with
education, the fear of Ebola is hampering access to services. Despite this
additional barrier, 60% of informants indicated that the situation has
actually become better compared to a year ago. This is due to the provision
of medicines, ambulances and increased awareness of the importance of
adequate healthcare among families. Additional support is still required,
with improvement of the health infrastructure mentioned by 30% of KIs as
a required intervention.
Income generation
According to KIs, only few could obtain sufficient income to adequately
provide for their families, which is in line with the available baseline data on
poverty levels within the county. When asked about the livelihoods
situation, 60% of KIs indicated that the situation is currently worse
compared to a year ago. The lack of job opportunities was identified as the
main concern before the outbreak. This problem has been compounded by
a general increase in unemployment and the closure of markets and
businesses.
ACAPS - ENAP Liberia April 2015
31
4. Grand Bassa
Key findings
The current situation in terms of access to education, livelihoods and health
is worse compared to a year ago, according to a majority of KIs surveyed.
The education situation is of particular concern, with KIs considering it
currently a higher priority than before the crisis. Fear of Ebola infection is
the main obstacle to accessing health and education, although the Ebola
incidence has been relatively low in the country and no cases have been
confirmed since December 2014. Districts 2 and 4 were those most in need,
and KIs agreed that this was mainly due to the limited number of health
facilities in these areas.
ACAPS - ENAP Liberia April 2015
32
Context
Palm oil and food crop production are the most important livelihood
activities in the county. The majority of citizens are engaged in informal
work in agriculture and / or petty trading (GoL, 2008).
Poverty rates are slightly above the average for rural Liberia, with 43% of
the population in the lowest quintile (DHS, 2013). School attendance rates in
Grand Bassa are the second lowest in the country, after Bong, at 23%. In
2008, a shortage of educational facilities was identified and available
schools were often overcrowded (GoL, 2008). Traditional culture remains
strong, with the Poro and Sande societies playing a major role in the
education and initiation of boys and girls and access to healthcare.
The first Ebola case in the county was reported mid-July. 54 confirmed
cases in total have been recorded, between July and mid-December (MoH,
04/2015).
Education
Access to education is currently seen as worse than before the crisis,
indicated by 80% of KIs. The inability of families to pay for transport and
materials is the main obstacle, both currently and before the crisis. The
poverty levels in the county partly explain this. In addition, the fear of Ebola
infection has further eroded the willingness of families to send their children
to school. Several schools have not yet reopened or staff are unavailable,
according to KIs.
Health
The fear of contracting Ebola is still the main obstacle to accessing
healthcare, according to KIs, although there has not been a confirmed
Ebola case in over five months. From all KIs surveyed nationwide, Grand
Bassa ranks the highest when it comes to the lack of trust in health facilities.
The majority of KIs judge the current situation to be worse than a year ago.
However, two KIs indicated that the health situation has improved, with
NGOs providing support to several health centres.
Income generation
The two main pre-crisis obstacles to income generation, a lack of
employment opportunities and limited skills and literacy, remain the largest
problems. KIs almost universally agreed that the ability of families to obtain
an income has decreased, as some have become unemployed due to the
impact of the Ebola crisis on the economy.
ACAPS - ENAP Liberia April 2015
33
5. Grand Cape Mount
Key findings
Access to education and livelihoods is still negatively impacted by the
consequences of the Ebola crisis, according to KIs. The results suggest
that the general decrease in livelihood opportunities has impacted the
financial means of families to send their children to school. There is a lack
of trust in the ability of health and education services to protect people from
Ebola transmission. KIs suggested increasing the provision of information
on health services and the importance of education to address this. Porkpa
district was identified as the district most in need, due to insufficient
schools, health facilities, roads and jobs.
ACAPS - ENAP Liberia April 2015
34
Context
Petty trading and agriculture are the main sources of income in the county.
Grand Cape Mount has the highest percentage of HHs dependent on
mining nationwide. 31% of 6–11 year olds are enrolled in primary school,
below the national average of 38% (DHS, 2013). Food insecurity is severe,
with just 21% of HHs being food secure, according to a 2012 assessment
(WFP, 06/2013). There is only one paved road in the county and accessibility to
some areas during the rainy season is virtually impossible due to damaged
bridges (GoL 2008).
Grand Cape Mount was one of areas with the highest Ebola incidence. The
first case was confirmed in July 2014, and cases continued to be reported
until mid-January 2015 (MoH, 04/2015). In October, roadblocks were set up to
limit travel to neighbouring counties and several communities in the county
capital, Robertsport, were put under quarantine (AFP, 24/10/2014; UNMEER,
30/11/2014).
Education
Access to primary education was ranked as a major problem, both before
the crisis and currently. 80% of the KIs indicated that access to education
is currently worse. Even before the crisis families did not have sufficient
resources to pay for transport and school materials, and access to
education has further diminished due to the fear of Ebola in schools. KIs
indicate a need to disseminate information on the importance of education,
which was highlighted as the main activity required to address existing
problems.
Health
The current main obstacles to accessing healthcare differ significantly from
a year ago, according to KIs. Before the outbreak, the main obstacles were
the lack of medicine, lack of financial resources and physical constraints.
The fear of Ebola and the lack of trust in health facilities have now
overtaken these. Almost half of KIs indicated that access to healthcare has
improved, with external actors currently providing support and medicine to
health facilities. However, KIs requested additional health supplies and
information on hospital services, to address the lack of trust in the health
system.
Income generation
The majority of KIs surveyed (over 80%), viewed the current income
generation situation as worse compared to a year ago. The lack of job
opportunities and limited literacy and skills are the main obstacles, as in the
rest of the country. Training to improve the skills is the most often
mentioned required intervention. Unemployment has increased in the
county due to Ebola, as not all companies have reopened or returned
productivity to pre-crisis levels.
ACAPS - ENAP Liberia April 2015
35
6. Grand Gedeh
Key findings
KI opinions differ widely on the current education, healthcare and income
generation situation, as compared to before the crisis. KIs who said that the
situation had improved assigned this to the increased support from the
government and NGOs. Hence, the variation in opinion could be an
indication of the localised differences in levels of support. Grand Gedeh is
one of nine counties where KIs highlighted the lack of payments to health
and education staff as a main obstacle. Gboe-Ploe, Tchien and Knobo were
singled out as districts where communities faced most problems, due to the
general lack of infrastructure. Girls under 18 and people with a disability
were highlighted as those most in need of support.
ACAPS - ENAP Liberia April 2015
36
Context
Grand Gedeh is mostly dependent on agriculture. Production is low, as land
cultivation is still relatively undeveloped. Consequently, people rely on
other sources of food and income, including hunting and petty trading (EMMA,
2012). Cross-border trade from neighbouring Guinea and Cote d’Ivoire is an
important source of food and non-food items. However, transport is
hampered by a lack of maintenance of roads and bridges, and
infrastructure damage inflicted during the civil war. The county is known to
have gold, diamond and iron ore deposits, but the natural resources are not
exploited (GoL, 2008). In December 2014, the county hosted over 20,000
refugees from Cote d’Ivoire, primarily in PTP refugee camp (LRRRC,
28/01/2015).
In 2013, 49% of children under 5 were chronically malnourished, the
highest rate in the country. This is primarily linked to feeding practices. Only
a third (33%) of newborns are breastfed within one hour of birth, compared
to the national average of 50%. 40% of children aged 6–11 attend school,
which is close to the national average of 38% and far above average for
the counties outside of Monrovia.
There have only been three confirmed Ebola cases, the first on 13
September and the last a month later, on 4 October 2014 (MoH, 04/2015).
Education
As in the rest of the country, the lack of financial resources is the main
constraint to accessing primary education. Despite the low number of
cases, Ebola infection was ranked by KIs as the second largest obstacle.
One KI highlighted that staff salaries had not been paid, while five others
stated that salary payments are one of the main interventions required.
Most KIs (around 80%) agreed that the situation is worse or similar, when
asked to compare current access to education to before the outbreak.
Health
The health situation has become worse, according to half of KIs. The fear
of Ebola is the main obstacle, followed by a lack of medicine. The payment
of staff salaries was also mentioned as an obstacle to accessing
healthcare, and a priority for intervention. NGO assistance and the
increased number of health staff has led to an improvement of the situation,
according to three KIs.
Income generation
As in the rest of the country, the lack of job opportunities continues to be
the main obstacle to income generation. This is exacerbated by the closure
/ disruption of markets and businesses, due to restrictions on cross-border
movement. However, half of KIs indicated that the current income
generation situation is largely similar to the pre-outbreak levels. NGO
support was mentioned by two KIs as having a positive impact, resulting in
better access to income sources compared to before the crisis.
ACAPS - ENAP Liberia April 2015
37
7. Grand Kru
Key findings
Health and education remain the two main priorities in the county. Income
generation has recently become more of a concern with markets closed or
disrupted, compounding the already low levels of employment. The needs
are primarily driven by the poor state of the road infrastructure, affecting
access to education, health and livelihoods. Grand Kru is the only county
where KIs ranked physical constraints as the main obstacle to accessing
health. Unlike in other areas, the quality of the staff and low number of
teachers is mentioned as one of the main concerns. Teacher training is
highlighted as the most important intervention required. Girls under 18 and
persons with a disability were main vulnerable groups highlighted by KIs.
ACAPS - ENAP Liberia April 2015
38
Context
Grand Kru is one of the most isolated and underdeveloped areas in the
country, with the highest poverty rates (DHS, 2013) and below average health
indicators. Families are largely dependent on subsistence farming and
there are limited other job opportunities (GoL, 2008). A 2013 assessment
found that 45% of people are food insecure, the highest proportion after
Bomi, compared to 18% nationally. (WFP, 06/2013)
The lack of infrastructure is particularly pressing. In 2008 it was estimated
that more than two thirds of the county was inaccessible by car (GoL, 2008).
As a result, the population face large difficulties in accessing basic services.
In 2013, 70% of women faced at least one problem in accessing healthcare.
50% indicated that this was due to a lack of money (46% nationwide) and /
or the large distances to the health facilities (40% nationwide) (DHS, 2013).
The county is prone to floods. At the end of 2014 floods affected rice,
cassava and palm oil production as well as cash crops (Ministry of Agriculture,
WFP, FAO, 30/11/2014).
Grand Kru was the last county affected by Ebola. The first case was
confirmed on 24 October 2014. Only three additional cases were reported
from the county, the last one in mid-November (MoH 04/2015).
Education
In 2008, most schools lacked proper structures and furniture (GoL, 2008).
Several interventions have taken place since then, but the low quality of the
school infrastructure is a continued obstacle highlighted by KIs. The low
quality and unavailability of staff is currently one of the largest problems.
Training education staff was one of the three main activities required, as
indicated by half of KIs. However, even if adequate education facilities were
in place, children would face large difficulties in accessing schools. The
main problem both before the crisis and currently is the ability of families to
pay for transport and education materials, due to the high levels of poverty
and remoteness of many populated areas. Six out of ten KIs said that the
current situation was worse than a year ago, with the fear of Ebola infection
in schools compounding existing access constraints.
Health
The county is relatively well covered by health infrastructure with 17 health
facilities, 2.5 per 10,000 people (Liberia Health System Assessment, 2015). The main
obstacles to access are physical constraints, followed by a lack of
medicine. Despite having the lowest incidence of Ebola, just four cases,
fear of Ebola and a lack of available staff in health facilities are the other
main current constraints. 60% of KIs judged the current situation as being
worse compared to a year ago, while 40% thought the situation was the
same. The lack of information on health care services was ranked as one
of the main concerns before the crisis, but KIs rank it as a much lower
priority now.
Income generation
The Ebola response in the county has resulted in additional, temporary, job
creation, but the overall situation has not structurally improved. The lack of
job opportunities is the main obstacle to making an income, followed by the
disruption of markets and businesses due to Ebola and physical / logistical
constraints.
ACAPS - ENAP Liberia April 2015
39
8. Lofa
Key findings
The Ebola outbreak in Liberia started in Lofa and the county reported one
of the highest number of Ebola cases. The peak of the outbreak took place
in the planting season between April and June. Although the county is
normally self-sufficient in terms of food, income generation and food are
currently the main priorities. The results indicate that the outbreak has
changed the behaviour of communities, with an apparent decreased
utilisation of traditional healers and avoidance of schools out of fear of
Ebola infection. Ebola orphans, survivors and their families were
highlighted as the groups most in need.
ACAPS - ENAP Liberia April 2015
40
Context
Lofa County was severely affected by the civil war. It suffered extensive
damage to infrastructure and basic social services, as well as mass
displacement. The county’s population is mostly reliant on agriculture, with
very limited activities in rubber, timber or mining (GoL, 2008). The county is
seen as the main breadbasket in Liberia, and is self-sufficient in rice
production (WFP, 06/2013). Most indicators regarding access to services are in
line with the national averages. Primary school attendance rates for
instance were at 36% of children aged 6–11 in 2013, compared to 38%
nationwide (DHS, 2013).
Liberia’s first confirmed Ebola cases were reported from Foya district in
Lofa, on 13 March. The county has been one of the worst affected during
the outbreak, with over 300 cases confirmed in the following nine months
(MoH, 04/2015; GoL, 21/10/2014; IMS, 02/04/2015). A Mercy Corps assessment in
October 2014 found that smaller harvests are expected for both staple and
cash crops, because of quarantined zones and restrictions on movement
in the county (Mercy Corps, 23/10/2014).
Education
Before the crisis, the main obstacle to accessing education was the inability
of families to afford transport, fees or materials. The fear of Ebola infection
in schools has overtaken these, and was ranked as the main problem now.
Almost 90% of the KIs indicated that access to education was worse than
a year ago. However, when asked about the main problems currently faced
by communities, KIs prioritised food, livelihoods, water, healthcare and
safety and security over education.
Health
The county has strong cultural and traditional practices. Before the crisis
families mostly decided to visit traditional healers, and KIs indicated that
this was the main obstacle to accessing formal health care. Currently, this
is no longer ranked as an obstacle. This could be an indication of an erosion
of trust of traditional healers, who were linked to localised outbreaks of
Ebola in the county, and the impacts of social mobilisation. The current
main problem is families’ fear of Ebola infection in health facilities. Most KIs
agree that access the healthcare is currently worse compared to before the
outbreak. The two KIs who judged the current situation as better,
highlighted the positive impact of increased awareness of the importance
of healthcare among communities.
Income generation
Before the crisis, there was no clear prioritisation of the problems in the
livelihood sector. The lack of job opportunities, limited literacy and limited
resources to invest in productive assets all ranked as major obstacles.
Currently, the three main problems are all a direct or indirect consequence
of the Ebola outbreak: the closure / disruption of markets, increased
unemployment, followed by a decrease in agricultural production. The
outbreak in the county peaked during the planting season, from April to
June, and the main harvest in October. Rice production is expected to have
decreased. This is due to the limited maintenance of the fields during the
growing season, because of quarantine measures and restrictions on group
work (Mercy Corps, 23/10/2014; WFP, 11/2014). The majority of KIs judged the
situation as worse compared to a year ago.
ACAPS - ENAP Liberia April 2015
41
9. Margibi
Key findings
Margibi County was one of the hotspots during the outbreak, with a large
number of Ebola cases reported in a short period of time. The main impact
of the Ebola crisis currently is the decrease in livelihood opportunities,
according to KIs. Although the fear of Ebola is an additional obstacle to
accessing healthcare and education, overall the current situation is similar
to that of a year ago. Girls under 18, Ebola orphans, survivors and their
families were highlighted as the groups most in need of support. Gibi district
was highlighted as the priority district, primarily due to a lack of roads and
job opportunities.
ACAPS - ENAP Liberia April 2015
42
Context
Margibi County is known for its rubber plantations, primarily the Firestone
and Salala plantations. These industries are an important source of income
and provide basic services including schools, shelter, and healthcare to
employees and families. Food crop production is not as widespread in
Margibi as in other counties, and it is a high deficit area in terms of rice and
cassava. Traders rely on suppliers from outside of the county for their stock
replenishment (Food Security Cluster, 13/11/2014). The healthcare system is below
average, with one of the lowest proportions of health workers for the total
population (Liberia Health System Assessment, 2015).
After Montserrado, Margibi recorded the highest number of Ebola cases at
392 confirmed cases. The first case was identified at the end of March and
cases were reported for the following 11 months (MoH, 04/2015).
Education
Attendance rates before the crisis were the highest of the counties outside
of Montserrado, at 40% of children aged 6–11 (DHS, 2013). Most KIs (70%)
indicated that the education situation has remained similar or has improved
compared to last year, mainly due to increased support from the
government and NGOs. However, as in other counties, the lack of
resources to pay for transport, fees or materials remains the main
constraint to accessing education. The fear of Ebola infection in schools is
currently one of the three main obstacles, which is unsurprising in a county
with one of the highest Ebola rates in the country.
Health
Margibi is one of the two counties where physical constraints are the main
obstacles to accessing healthcare, treatment or preventive care. The lack
of medicine at health centres is an additional main obstacle. Before the
crisis visiting traditional healers, instead of formal healthcare centres, was
one of the main obstacles identified by KIs. This is no longer mentioned as
a concern, which could indicate that communities have lost trust in
traditional health practices due to the Ebola outbreak and subsequent
social mobilisation. 70% of KIs indicated that the current situation is similar
or better than last year, which could indicate that most of the current
problems were already of concern before the outbreak. Four KIs indicated
that the situation is currently better, due to NGO support and the increased
awareness among communities of the importance of health care.
Income generation
The main livelihood activities are rubber tapping, charcoal production and
subsistence agriculture. KIs mention increase unemployment as one of the
main concerns, partly due to Ebola related closures and the disruption of
markets and businesses. The rubber plantations are highly dependent on
transport, and Ebola related border closures and travel restrictions
throughout the region have severely limited the import and export of goods.
Most of these restrictions have now been lifted, but KI responses indicate
that the market has not yet fully recovered. 69% of respondents indicated
that the current situation is still worse than before the outbreak.
ACAPS - ENAP Liberia April 2015
43
10. Maryland
Key findings
The main problems mentioned related to access to education, healthcare
and livelihoods. Most obstacles to accessing basic needs were already of
concern before the crisis, including a lack of jobs and physical constraints.
The exception is the fear of Ebola, which prevents families from sending
their children to school. Surprisingly, access to water was the highest
priority before the crisis, while KIs now consider it one of the lowest. As the
water infrastructure has not been greatly improved over the last year, it is
likely that this difference in ranking stems from the relative increased
importance of income generation and health. Physical constraints, primarily
the poor roads, are one of the main reasons why families cannot reach
basic services. Maryland experiences a long rainy season, during which
roads are mostly inaccessible.
ACAPS - ENAP Liberia April 2015
44
Context
Agriculture is the main source of income for Maryland, and rubber
cultivation provided an income for around 40% of HHs in 2008. The county
borders Cote d’Ivoire and border communities, including the capital Harper,
rely on cross-border trade, especially during the rainy season. As in the rest
of the southeastern region, the road infrastructure in Maryland is
particularly poor, especially during the prolonged rainy season (GoL, 2008).
The county hosts at least 10,000 refugees who fled the 2010 post-election
violence in Cote d’Ivoire (LRRC 28/10/2015).
In 2013, household wealth was above average when compared to the other
counties in the southeastern region. Only 21% of HHs were categorised as
part of the poorest 20% in the country, compared to 37% for the rest of the
southeastern region. School attendance rates are above the national
average, with 39% of children aged 6–11 attending primary school in
Maryland in 2013 (DHS, 2013).
The first Ebola case in Maryland was recorded at the start of September
2014. Four confirmed cases were recorded in total (MoH, 04/2015).
Education
KIs disagree on how the current education situation compares to before the
crisis. Half stated that the situation was worse, the other half judged it to be
better. The main reason for this improvement is related to the behaviour of
students and parents – after a long period without access to education due
to school closures, families value schooling more, according to KIs. There
is more agreement on the current main problems, with the lack of financial
resources identified as the main obstacle by all informants surveyed,
followed by the fear of Ebola infection in schools.
Health
The lack of information on healthcare services was a main concern before
the crisis, but no KIs highlighted this as a current issue. This indicates that
the Ebola outbreak and response has led to increased awareness of
available services. Unlike the rest of the counties, fear of Ebola was not
highlighted as a main concern. The lack of medicines and physical
constraints, including the poor state of the road infrastructure, are currently
the main obstacles to accessing healthcare. One third of the KIs, however,
consider the current situation to be better than before the outbreak. KIs
attribute this improvement to the additional support of NGOs and greater
awareness of health services among communities.
Income generation
It was widely assumed that the border closures would have a major impact
on the livelihoods of communities in Maryland, which are partly reliant on
trade with Cote D’Ivoire. However, the main problems mentioned by KIs
were the lack of job opportunities and physical constraints, which were the
main problems before the crisis. KIs did not mention problems which were
directly related to the Ebola crisis. Several mentioned the positive impact
of income generating opportunities created by NGOs providing services in
the county.
ACAPS - ENAP Liberia April 2015
45
11. Montserrado
Key findings
Despite the concentration of services in Montserrado, the current problems
expressed by KIs are surprisingly similar to those faced by rural
communities. This is because the majority of problems are access related.
Although education and health facilities are largely available in most parts
of Montserrado, families do not have the financial means to access
services. The existing access problems are compounded by the fear of
Ebola, which impacts access to education and health. The suggested
activities required are water, health and road infrastructure improvements,
indicating that service availability remains of concern.
ACAPS - ENAP Liberia April 2015
46
Context
Monrovia is the centre of most political, economic and social activity in
Liberia. It has higher standards of education, health, security and
infrastructure. Illustratively, 50% of the population in the county was part of
the richest 20% in Liberia in 2013, and almost all HHs were considered to
have acceptable food consumption. School attendance rates are far above
average, with 53% of children aged 6–11 attending primary school
compared to 38% nationwide. However, large scale urban migration is
straining the capital’s limited infrastructure. More than 20% of rural HHs
having at least one family member who has migrated to Monrovia,
according to a 2013 study (DHS, 2013; WFP, 2013).
The situation in rural Montserrado is comparable or worse, when compared
to other rural counties. In the district St Paul River for instance, there is only
one paved road. A WFP assessment in 2013 found diarrhoea rates in rural
parts of Montserrado at 25%, far above the national average of 17% (GoL,
2008, WFP, 06/2013).
The first Ebola cases in Monrovia were reported in mid-May (MoH, 04/2015).
The city was unable to cope with the high number of infections that followed
in June, and a state of emergency was proclaimed at the start of August.
At the end of August, the government quarantined the city’s West Point
slum, hosting over 75,000 people, in an attempt to slow the spread.
Violence between the population and armed forces broke out, resulting in
the death of one teenager (WHO, 01/2015).
Education
The lack of financial resources was the main obstacle to accessing
education, and still is. Over 70% of the KIs indicated that access to
education is currently worse than before the crisis. One of the main
additional obstacles is the fear of Ebola, which discourages families from
sending their children to school. The third main obstacle is the need for
children to support their families, with children forced to help with domestic
chores or work outside of the house.
Health
The lack of financial resources is the main current obstacle to accessing
healthcare. 200 of the 240 health facilities in the county are private, which
are more expensive than their government counterparts (Liberia Health System
Assessment, 2015). The fear of Ebola infection at health centres is the one of the
main concerns. The lack of trust in the health system is currently the third
largest obstacle to accessing healthcare, and was a major problem before
the crisis. 60% of KIs indicated that access to healthcare had worsened
while over 20% indicated that it had improved, primarily due to the
increased support to healthcare centres.
Income generation
The lack of job opportunities and lack of skills are currently the main
obstacles to income generation, as they were before the crisis. However,
the majority of KIs view the current situation as worse, compared to before
the crisis. The jobs lost during the Ebola crisis have not all been replaced,
and the increase in unemployment is still a major concern.
ACAPS - ENAP Liberia April 2015
47
12. Nimba
Key findings
The education, health and income generation situation in Nimba is currently
worse than a year ago, according to KIs. The deterioration can be traced
to the increased difficulties in accessing healthcare and education, due to
the fear of Ebola infection. Livelihood opportunities remain affected by the
temporary border closures and movement restrictions, with Ebola related
unemployment highlighted as one of the main concerns. Gbi, Doru,
Yarwein, Mehnsonnon and Kparblee were emphasised as the main districts
in need of support, due to the limited number of schools, health facilities
and poor state of the infrastructure.
ACAPS - ENAP Liberia April 2015
48
Context
Nimba is the second most populated county in Liberia. During the civil war
the county saw some of the heaviest fighting, resulting in large scale
displacement and widespread destruction of infrastructure. The exploitation
of mineral resources was an important source of revenue and jobs before
the conflict. Now it is mostly limited to the illicit mining of gold and diamonds,
except for the mining by ArcelorMittal which is extracting iron ore in the
county (GoL, 2008). Over 5,000 Ivorian refugees, of the 38,000 who fled to
Liberia following the 2010 post electoral crisis, reside in Bahn camp in
Nimba (LRRRC, 28/01/2015; UNHCR, 23/10/2014). In 2013, school attendance rates
among 6–11 year olds were the third highest in the country, after
Montserrado and Margibi, at 34% (DHS, 2013).
The first Ebola case in Nimba was on 4 July, and cases continued to be
reported until mid-December (MoH, 04/2015). A Mercy Corps assessment in
October 2014 found that one of the main consequences of the outbreak
was the halt in cross border trade. With the shift from importing goods
overland from Guinea and Ivory Coast to shipping through Monrovia, HH
costs have increased significantly while income opportunities have
declined (Mercy Corps, 04/11/2014).
Education
13 out of 15 KIs indicated that the current situation has deteriorated
compared to a year ago. Access to education was ranked as the main
current need. The main obstacles to accessing education now include the
lack of resources for families and the fear of Ebola infection. In addition,
families do not always see the added value of
education. The need for social mobilisation is
one of the main activities required within the
county.
Health
Fear of Ebola infection and the lack of money for transport and services are
currently the main obstacles to healthcare, according to the respondents.
KIs proposed the training of (additional) health staff as an intervention to
address existing obstacles to healthcare. Unlike other counties, the lack of
medicine was not considered as a major obstacle to accessing healthcare
in Nimba. While it was common for families to visit traditional healers before
the crisis, this behaviour was no longer mentioned as a predominant
obstacle to accessing conventional healthcare. Four out of 15 KIs indicated
that the health situation was better than the same time last year, as there
is more awareness among communities on the importance of healthcare
and the services available.
Income generation
11 out of 15 KIs considered the current livelihoods situation as worse
compared to a year ago. Income generation is ranked as a higher need
than a year ago. Before the outbreak, the limited literacy and skills and lack
of job opportunities were the main obstacles to obtaining an adequate
income. These issues remain the main concerns. Training to improve the
skills of families was one of the main interventions proposed by KIs. In
addition, KIs indicated that there is currently an increase in Ebola related
unemployment, most likely partly caused by the border closures which have
affected trade with Guinea and Sierra Leone.
ACAPS - ENAP Liberia April 2015
49
13. River Gee
Key findings
Education and health are the main priorities within the county, according to
KIs, with the aftermath of the Ebola outbreak continuing to affect access to
both. Although there were only a limited number of cases, the fear of Ebola
continues to be a reason why families have difficulty accessing treatment
at health centres or sending their children to school. Transport to health
facilities is a major concern in the remote parts of the county, and pregnant
women were highlighted as the group most in need of support. KIs
specifically mentioned the need for additional health supplies, with half of
mentioning this as one of the main interventions required. Glaro, Gbeapo
and Nyenebo are cited as priority districts, primarily due to the lack of health
and education staff.
ACAPS - ENAP Liberia April 2015
50
Context
River Gee has one of Liberia’s lowest population densities. While road
access is limited throughout rural Liberia, River Gee is generally considered
as particularly remote (Liberia Health System Assessment, 2015). Agricultural
productivity is mainly focused around subsistence agriculture and
communities have very limited access to markets (GoL, 2008). Poverty levels
are high, with 73% of the population belonging in the poorest 40% of
Liberia. 38% of children aged 6–11 attend school, equal to the national
average (DHS, 2013).
The county has seen only a small number of Ebola cases. The first Ebola
case was recorded at the start of August 2014, and the last case two
months later (MoH, 04/2015).
Education
The lack of financial resources to pay for transport and materials was
ranked as the main education obstacle in River Gee. However, the situation
in the county differs from other areas when it comes to awareness on the
importance of education. Families or students do not see the added value
of education, which ranked as one of the main obstacles both currently and
before the crisis. 64% of the KIs ranked the current situation as similar or
better than before the Ebola outbreak. This is primarily due to the positive
impact of the back to school campaign on families’ perceptions on the
importance of education.
Health
Fear of Ebola is currently the main obstacle to accessing healthcare,
followed by the lack of medicine. The provision of health supplies was
highlighted as one of the main interventions required. In addition, River Gee
is one of the six counties where the lack of payment of staff salaries was
mentioned as one of the main obstacles. KIs are divided on how the current
situation compares to the pre-crisis conditions, with 45% judging the
situation as worse and 55% as similar or better.
Income generation
The lack of job opportunities is by far the most important obstacle to
accessing an income currently. According to KIs, few people could earn
sufficient income to adequately provide for their families before the crisis
and the situation has worsened. A decrease in agricultural production, the
main source of food for most communities, was highlighted as a main
concern as a result of the Ebola outbreak. The other two main concerns
are more structural, with the lack of job opportunities and physical
constraints already a large concern before the crisis.
ACAPS - ENAP Liberia April 2015
51
14. Rivercess
Key findings
According KIs, the current priorities are the same as before the crisis –
income generation, followed by education and water. However, the
intensity of the problems has increased, with most KIs agreeing that
education, livelihoods and health situation is currently worse. The
underlying cause of most of the problems is the lack of income generating
opportunities and basic infrastructure. A large variety of interventions were
proposed, which could mean that a wide range of different interventions are
required in this isolated and highly underdeveloped county. The group most
in need of support are girls under 18, according to KIs.
ACAPS - ENAP Liberia April 2015
52
Context
Rivercess is one of the poorest counties in Liberia, with over 70% of the
population being among the poorest 20% in country (DHS, 2013; WB, 2011). Low
levels of education partly explain this, with over 60% of the female
population having no formal education, compared to 47% countrywide. (DHS, 2013).
The main livelihood activities today include palm oil production, hunting,
food crop production and fishing. Gold and diamonds are currently mined
illegally, and only at the artisanal level. 80% of HHs do not have access to
safe drinking water and 96% have no toilet facilities. School attendance
rates are among the lowest in the country, 24% of children aged 6–11.
These very low indicators could partly be explained by the geographical
isolation of Rivercess, where road access is among worst in the country. (GoL 2008, WFP 06/2013)
The first confirmed Ebola case in the county was on 8 July, three months
after the start of the outbreak in Liberia, and cases continued to be reported
until 25 November.
Education
The education situation has worsened compared to the same time last
year, according to over 90% of KIs. The decrease in income sources has
made it more difficult for families to afford transport and school materials, a
main concern currently and before the crisis. KIs indicated that there were
only a limited number of schools available before the crisis, and physical
constraints, including long distances, make it difficult for children to access
them. Due to the high levels of poverty, the need for children to support the
family was a significant obstacle before the crisis and currently.
Health
The health situation has become worse compared to before the outbreak,
according to over half of KIs. Rivercess is the only county where visiting
traditional healers has become a more prominent concern. This was ranked
as a major problem before the crisis and is currently the main obstacle to
accessing conventional healthcare.
Income generation
KIs indicated that before the crisis, only a few people earned sufficient
income to provide for their families, and the majority agreed that the current
situation is similar or worse. The lack of job opportunities is the main
obstacle and, according to KIs, the Ebola outbreak has led to increased
unemployment. Limited literacy and skills is the second main obstacle, and
skills training was one of the main activities proposed.
ACAPS - ENAP Liberia April 2015
53
15. Sinoe
Key informants
KIs largely agree that the current education, health and income generation
situation is worse than before the crisis. The current priority concerns are
income generation, education and food assistance. As in the rest of the
southeastern counties, logistical constrains in accessing services are one
of the main obstacles and improvement of roads was the main intervention
recommended. Butaw, Greenville, Jeadepo and Jaedae were highlighted
by KIs as the districts most in need of support, primarily due to the general
lack of infrastructure.
Context
Communities in Sinoe rely primarily on agriculture, fishing and mining
activities. Despite the relatively high number of health centres, people are
still forced to travel long distances to receive treatment. In 2011, only 14%
of the population lived less than 5km from health facilities (Health Facilities
Assessment, 2015). 33% of children aged 6–11 attended school, above average
for rural Liberia (DHS, 2013). A small number of Ebola cases were reported in
Sinoe from mid-August to end of December (MoH, 04/2015).
Education
The main obstacle to accessing education is the lack of financial resources,
currently and before the crisis. The other main obstacle is particular to
Sinoe County. Since the population is spread out over a large territory,
people have to travel long distances to reach services. As a result, physical
and logistical constraints were among the main obstacles before the crisis,
followed by the limited number of schools available. These issues remain.
In addition, the fear of Ebola infection has further decreased access. Over
80% of KIs viewed the current situation as worse compared than before the
crisis.
Health
Before the crisis and currently, the main obstacle was that families chose
to visit traditional healers. The fear of Ebola is also a major obstacle at the
moment. Difficulty reaching health facilities was mentioned by KIs, with
physical constraints ranked as the third main concern. The vast majority of
KIs (84%) viewed the current situation as worse than before the crisis.
However, unlike in other counties, KIs do not currently rank health as one
of the highest priorities.
Income Generation
Current access to livelihoods is worse compared to the situation in April
2014, according to over 80% of KIs. The lack of job opportunities was the
main concern before the crisis, and KIs agree that the current situation is
either similar or worse. The Ebola outbreak still impacts access to
livelihoods, due to higher unemployed levels and the closure and disruption
of markets.
ENAP Liberia April 2015
Annex A - Questionnaire
KI QUESTIONNAIRE
ACAPS/Building Markets – Multi Sector Needs Assessment
County:
Date (dd/mm/yy):
Enumerator name:
Name Key Informant:
Phone Number Key Informant:
Time start interview
INTRODUCTION
Good morning/afternoon. My name is..................... I am part of a team from ACAPS/Building Markets. Our
organization assesses the current needs in Liberia. We would be really grateful if you could spare some
time to answer a few questions. The objective of the survey is to have a better understanding of the main
needs and challenges faced by the people of Liberia. The questionnaire covers education, health and
income generation and takes about 30 minutes. To thank you for your time, you will receive USD 2 in
airtime once you have finished the complete questionnaire.
The information you provide will be treated with total confidentiality and your name will not be reflected
in the final report. It is within your rights to choose whether or not to take part or, at any point, to
withdraw, but it will really be helpful if you could spare time to talk to us.
We will ask you about the situation in your county. There are no ‘right’ or ‘wrong’ answers to the
questions presented but honest and sincere answers are required. It is OK if you do not know the
answer to a question.
If you are willing to participate we will ask you several questions to see if you are the right person to talk
to. Would you like to participate in this survey?
□ Yes □ No
I.ELIGIBILITY CRITERIA
I.I County □ Correct □ Incorrect
I.II What is your position within your county and what is your
engagement with communities in your county?
□ Correct □ Incorrect
I.III What is the name of the senator in your county? □ Correct □ Incorrect
I.IV What are the main sources of income within your county? □ Agriculture □ Trade and Services
□ Fishing □ Mining
□ Casual Labour □ Rubber
□ Charcoal □ Other__________
I.V Response to source of income □ Correct □ Incorrect
I.VI In your opinion, how knowledgeable are you about the
situation of communities in your county? Do you know about the
situation in the whole county?
□ Yes □ No
□ NOT ELIGIBLE
From your answers to these questions I understand that we are looking for a different person to respond to our
questionnaire. So the survey stops here. Thank you for your time.
□ ELIGIBLE
You are the right the person to talk to. I would now like to proceed with the full questionnaire, which will take
about 30 min. Do you have time right now to answer these questions or shall I call you back later?
□ Continue with full questionnaire □ Scheduled for a later date
ACAPS - ENAP Liberia April 2015
56
A.RESPONDENT CHARACTERISTICS
LET’S START WITH SOME GENERAL QUESTIONS ON YOUR BACKGROUND
A1. What is your gender and age? Male Female Age
A2. Where do you live? County District Town
A3. What were you doing for a living over the past 12 months? (multiple answers possible)
Government employee (excl. teacher, health staff)
Teacher
Health Staff
Other services
Clerical
Petty trade/Business
Domestic service
Agriculture
NGO
Media
Other (Please specify)_____________________
A4. What is the highest education level that you have completed? (choose only one answer)
Never been to school
Primary school
Secondary school
University
Technical vocational school
Other (Please specify)_____________________
A5. In case this line is cut, do you have another
number where we can reach you?
B. EDUCATION
B1. How would you describe the education situation before the outbreak, at this time last year?
Most children went to primary
school
Few children went to
primary school
Almost no children went to
primary school Do not know
B2. For children not going to primary school, what were the 3 main obstacles at this time last year?
B3. Of these 3…
1. What was the biggest problem? ________________
2. What was the 2nd biggest problem? ______________
3. What was the 3rd biggest problem?_____________
B4 Now let’s talk about the current situation. How would you describe the current education situation as
compared to this time last year?
Worse Similar Better Don’t know
B4.1 IF BETTER, why?
B5. Right now what are the 3 biggest problems?
B6. Of these 3
1. What is the biggest problem? ________________
2. What is the 2nd biggest problem? ______________
3. What is the 3rd biggest problem?_____________
ACAPS - ENAP Liberia April 2015
57
C. HEALTH
C1. How would you describe the health situation before the outbreak, at this time last year?
Most people went to a health
facility to get treatment
Few people went to a
health facility to get
treatment
Almost no one went to a health
facility to get treatment Do not know
C2. For people that did not go to a health facility to get treatment, what were the 3 main obstacles at this
time last year?
C3. Of these 3…
1. What was the biggest problem? ________________
2. What was the 2nd biggest problem? ______________
3. Which one was the 3rd biggest problem?_____________
C4. Now let’s talk about the current situation. How would you describe health situation currently as
compared to this time last year?
Worse Similar Better Don’t know
C4.1 IF BETTER, why? (if worse/similar/don’t know move to question C5)
C5. Right now what are the 3 biggest problems?
C.6. Of these 3
1. What is the biggest problem? ________________
2. What is the 2nd biggest problem? ______________
3. What is the 3rd biggest problem?_____________
D. INCOME GENERATION
D1. How would you describe the means for communities to generate an income before the crisis, at this
time last year?
Most people could access
sufficient income to
adequately provide for their
families
Few people could access
sufficient income to
adequately provide for their
families
Almost no one could access
sufficient income to
adequately provide for their
families
Do not know
D2. For people that could not access sufficient income to adequately provide for their family, what were
the main obstacles at this time last year?
D3. Of these 3…
1. What was the biggest problem? ________________
2. What was the 2nd biggest problem? ______________
3. Which one was the 3rd biggest problem?_____________
D4. Now let’s talk about the current situation. How would you describe the possibility for families to
generate an income currently compared to this time last year?
Worse Similar Better Don’t know
D4.1 IF BETTER, why? (if worse/similar/don’t know move to question D5)
D5. Right now what are the 3 biggest problems?
ACAPS - ENAP Liberia April 2015
58
E. NEEDS PRIORITISATION
E1. You mentioned problems in healthcare, education and making an income. Before the outbreak of
Ebola there were also problems with water, safety and security and accessing food. Among those 6, at
this time last year,
1. What was the biggest problem? ________________
2. What was the 2nd biggest problem? ______________
3. Which one was the 3rd biggest problem?_____________
E2. Let’s talk about now, what are the top 3 problems for your county currently?
1. What is the biggest problem? ________________
2. What is the 2nd biggest problem? ______________
3. Which one is the 3rd biggest?_____________
E3. In your opinion, what are the 3 main activities needed in your county to address these problems right
now?
F. PRIORITY DISTRICTS
F1. In which districts in your county do people currently face the most problems?
Districts Other (specify)_________ Do not know
F2. And why do these districts face more problems than other districts?
G. VULNERABLE GROUPS
I am going to read a list of 10 groups. Which 3 groups are in most need of support? (READ OPTIONS OUT
LOUD choose only 3 answers )
Girls under 18
Boys under 18
Older persons (60 and above)
Pregnant women/lactating women
Persons with disability
Female head of household
Child head of household
EBOLA orphans
EBOLA survivors and their families
EBOLA workers and their families
Do not know/No response
Other (specify)____________________
H. OTHER
Is there anything else that you want to mention about how the current situation is different from the same
time last year?
CLOSURE
D6. Of these 3
1. What is the biggest problem? ________________
2. What is the 2nd biggest problem? ______________
3. What is the 3rd biggest problem?_____________
ACAPS - ENAP Liberia April 2015
59
We greatly appreciate your time and participation in this survey. Could we contact you again if we were to need
further clarification or for future monitoring surveys?
Yes No
FOR RESEARHCERS ONLY: On a scale of 1 to 3, how much confidence do you have in the accuracy of
the responses?
1 Answers provided appear consistent / no reason to doubt accuracy of responses
2 Respondent might not have understood all questions /some reason to doubt responses
3 Respondent clearly did not understand most questions/ significant reason to doubt responses provided
FOR RESEARCHER ONLY: COMMENTS
Any comments or remarks, comments on confidence
Status survey: Completed Partly completed Not eligible
Duration survey (in
minutes)
ENAP Liberia April 2015
Annex B – Methodological annex
Aldo Benini - Survey estimation and the measurement of priorities A note for ACAPS/Building Markets 23 April 2015
Summary
The Liberia Multi Sector Needs Assessment is a sample survey of key informants polled on changing sector priorities
between spring 2014 - before the Ebola crisis - and spring 2015 - the beginning of the recovery period. As such, its
findings are subject to sampling variance. This note discusses sector priorities when the uncertainty due to sampling
is taken into account.
Priorities were measured with a device known as the "Borda count". The note gives background for this measure,
discusses its assumptions and addresses a specific challenge. The challenge results from the fact that key informants
were made to prioritize six sectors. Two of these were substantively closely related - "access to food" vs. "income
generation". If these two are treated separately, it appears that currently income generation is of similar high priority
with education and health. However, when they are suitably combined, "food/income" dominates all other sectors in
current priority.
The focus of this note is on country-wide statistics, for which survey estimation is essential in order to avoid bias.
Many users will be interested as much or more in district-specific results ("counties" in Liberia - there are 15). For
reasons particular to this key informant sample, weighting within given counties is not needed. However, we present
county-wise estimates of one Borda count variable to demonstrate that for most counties the difference vis-à-vis the
national average is not significant. This cautions against interpreting small differences in local priorities as a basis for
different recovery policies. Policy should focus on robust differences.
Introduction
The Liberia Multi Sector Needs Assessment (henceforth "the Assessment") elicited opinions from key informants
selected in each of the 15 counties (districts) in the country. The purpose of the survey was to establish challenges
that people were facing in transiting from the Ebola crisis into the recovery phase. The survey also sought to establish
relative priorities of unmet needs across several sectors.
The key informants were asked to single out the most important, second most important and third most important
elements in sets of challenges. A set would typically relate to a specific sector; the Assessment addressed three
sectors: education, health, income generation. The questions about challenges were asked twice, once
retrospectively concerning conditions in spring 2014 (before the Ebola crisis), and then referring to the current
situation (March - April 2015). In this segment, Questions B2 and B3 illustrate the elicitation process by the example
of education before the crisis.
ACAPS - ENAP Liberia April 2015
61
Figure 1: Most important challenges - example of the elicitation process
In the inter-sectoral perspective, the interviewers asked the key informants to locate the three biggest problems in
six sectors, retrospectively and for the current time. In addition to education, health, income generation, the questions
referred also to water, safety and security, and access to food. These questions produced priority rankings for six
sectors in two periods.
All in all, Assessment workers interviewed 216 key informants. All interviews were by phone. The key informants
were selected randomly from lists, separate for each county, of government officials, NGO and media workers, as
well as traditional leaders. They underwent a simple test at the start of the interview to ascertain that they were
sufficiently knowledgeable of the needs of people in their respective counties. Of the 330 informants that the
Assessment approached, eventually 216 completed the interview.
Survey estimation As in most key informant surveys, the sample is not a proper probability sample. The normal assumptions of survey
estimation do not hold: There is no sampling frame that covers the target population (the set of key informant-capable
persons) to a high degree; for the actual sample members, the selection probabilities are not known4. Therefore,
parameters of interests - means, proportions - can be estimated under far-reaching idealizations only:
By background, the sampled key informants are clearly not representative of the general population, and
plausibly not of those residents capable of answering the interview questions either (see the breakdown in
other sections of the methodological appendix). Yet, we assume that the views of those interviewed in a
given county tend towards the true values of the population parameters. In other words, if in county X a
majority of the households perceived the unmet need for health care to be stronger than the unmet need for
water, AND the interviews could be repeated with other informants of similar background infinitely many
times, a preference for health care support would emerge.
4 This is not a problem of this Assessment only or of surveys in Liberia. It has been observed for expert elicitation studies in general; see e.g., "it is difficult to recruit a valid probability sample of experts because of difficulties in (a) defining the universe from which a sample should be drawn and (b) overcoming selection biases associated with experts’ availability and willingness to participate" (Swackhamer and Hammit 2010: Annex, page 6).
B. EDUCATION NOW, I WOULD LIKE TO ASK YOU ABOUT ACCESS TO PRIMARY EDUCATION (BELOW GRADE 6) WITHIN YOUR COUNTY. FIRST, WE WILL DISCUSS THE SITUATION BEFORE THE EBOLA CRISIS, AFTERWARDS WE WILL DISCUSS THE CURRENT SITUATION. B1. How would you describe the education situation before the outbreak, at this time last year? I will give you 3 options (READ OPTIONS OUT LOUD choose only one answer ) Most children went to primary school
Few children went to primary school
Almost no children went to primary school
Do not know
B2. For children not going to primary school, what were the 3 main obstacles at this time last year? (Only read out below options if respondent has difficulties in answering the question) No or limited number of schools available Low quality of teachers and school staff not available Low quality infrastructure (e.g. buildings, furniture) Lack of adapted infrastructure and service for children with disabilities Schools not perceived as safe because of risks, incl. corporal punishment.
Families did not see the added value of education Children needed to help the family (incl. domestic chores, child labour) Families did not have money to pay for transport, fees or books and school uniform or materials No feeding programme in schools
Physical and logistic constraints to access education (roads damaged, no transport, etc.) No problem Do not know/No response Other (Please specify)_____________
B3. Of these 3…
1. What was the biggest problem? ________________ 2. What was the 2
nd biggest problem? ______________
3. What was the 3rd
biggest problem?_____________
ACAPS - ENAP Liberia April 2015
62
The practical consequence is that within the county every interviewee can be thought of as representing the same
number of residents. There is no basis to stipulate that some interviewees represented more persons, and others
fewer.
Although the views of key informants plausibly were influenced by what they had learned about conditions
outside their counties - particularly those who had family elsewhere -, we assume that every one spoke
strictly to conditions in the county of residence or official posting.
The practical consequence is unequal weighting across counties. The number of persons that a key informant
represents is not simply the national population divided by the sample size. It is the county population divided by the
number of key informants interviewed about the county.
Accepting these idealizations, survey estimation is needed for two reasons:
The number of persons represented by a key informant varies considerably across counties. Therefore
national estimates differ depending on whether they are weighted or not.
These differences turned out to be relatively minor, but this was not known beforehand. For example, on the priority
measure for the current unmet need for water (explained further below), the unweighted sample mean was 0.81, the
design-weighted mean was 0.94. Most other estimates showed smaller differences.
Since every key informant within a given county is thought of as representing the same number of persons,
weighted and unweighted county estimates are identical. However, county samples are relatively small, and
therefore estimates are highly uncertain.
These uncertainties can be expressed in confidence intervals, ranges extending to both sides of the estimated
parameter (mean, proportion) into which the true values fall with a high probability. Confidence intervals are important
as safeguards against attributing importance to local differences that may only reflect sampling variance. For
example, there was considerable variation in the priorities expressed of the unmet need for water. However, among
the fifteen counties, only in one (in Bong County) did the key informants concur on such a high priority that the
resulting score exceeded the national mean with strong statistical significance. Neglecting the uncertainty might
cause the assessment users to divert sector resources to counties with spuriously high priorities.
ACAPS - ENAP Liberia April 2015
63
Figure 2: Example of priority estimates, by county, with confidence intervals
In this example, it is obvious that water was a much higher priority than the country-wide mean score for the key informants in one of the 15 counties only.
Sample, sampling weights and estimation set-up The persons represented by one key informant are used as sampling weights, as detailed in this table, which also compares planned and actual samples. Table 1: Key informants interviewed and sampling weights, by county
County Poulation
2014 (projected)
Key informants
Planned Actual
Persons represented by one informant
Bomi 106,146 15 15 7,076
Bong 420,436 15 15 28,029
Gbarpolu 108,382 15 14 7,742
Grand Bassa 281,276 15 14 20,091 Grand Cape Mount 155,384 15 11 14,126
Grand Gedeh 158,575 15 14 11,327
Grand Kru 72,459 10 10 7,246
Lofa 346,430 15 17 20,378
Margibi 263,696 15 13 20,284
Clearly, Bong County issignificantly higherthan the national mean (maroon)
0.5
11.5
22.5
Prio
rity
me
asu
re
Bom
i
Bon
g
Gb
arp
olu
Gra
nd
Bassa
Gra
nd
Cap
e M
ou
nt
Gra
nd
Ge
de
h
Gra
nd
Kru
Lo
faM
arg
ibi
Ma
ryla
nd
Mo
nts
err
ado
Nim
ba
Riv
er
Gee
Riv
erc
ess
Sin
oe
County
Note: 216 key informants total. Blue dots: county means; blue spikes: 95% confidence intervals.Maroon lines: National mean (thick), 95%CI bounds (thin).
Water as a major problem (before the Ebola crisis)
ACAPS - ENAP Liberia April 2015
64
While the weights vary by a factor of six, the size of the standard errors generally varies by much less. The sample was stratified on the counties, with the observations as the primary sampling units and their weights as detailed above5. A finite population correction was not made6.
Measurement of priorities
Borda counts In the interview, priorities were elicited as "top three problems". This figure reproduces the elicitation wording for the needs prioritization across six sectors. Figure 3: Elicitation of inter-sectoral priorities
Subsequently, the stated priorities were scored for "modified Borda count" calculations. The sector taken as the highest priority received a score of 3, the second-highest a score of 2, the third 1. The non-priority sectors were all scored 0. Borda counts (Borda 1781; Wikipedia 2011; Benini 2013) are preference models in which items (candidates in an election, sectors in a needs assessment, etc.) are assigned votes in descending order of preference. In the original scheme, meant for public elections, the voter gives his top candidate a number of votes equal to the number of candidates minus one (N - 1), and his i-th preferred candidate N - i votes. The preferences of the voters (electors or, in our case, key informants) in the constituency are aggregated by counting the votes of each candidate. These Borda counts are thus ratio-level variables, with meaningful intervals and zero points. Modified Borda counts - to be precise: the kind of modification relevant here - restrict the voting to M < N choices, with M votes given to the first option, M - 1 to the second, etc., and zero votes to the N - M least favored candidates. The resulting counts certainly have an interval-level interpretation. The zero point may or may not have a substantive interpretation, depending on context and purpose, but it may not be relevant to establishing an informative preference order either (think of temperature measurements for comparisons between points in space and time, without regard for the absolute zero).
5 As often, the gain in precision from stratification is small. For example, the estimate of the mean of the priority measure for water for the nation currently has a standard error of 0.093 neglecting stratification, vs. one of 0.088 taking it into account. The design effects (DEFF) - the ratios of variances under our sample design to those under an equal-size simple-random sample - are 1.47, respectively 1.23. But a non-stratified sampling would not have been acceptable due to the risk of ending up with even greater imbalances of over- and under-sampled counties. 6 We mention this because in a similar assessment recently conducted in Sierra Leone, strata were formed of districts, based on exposure to the epidemic and agricultural livelihoods dependency. As primary sampling units, districts were drawn from the small number of districts in the stratum; thus FPC was appropriate, improving precision.
E. NEEDS PRIORITISATION E1. You mentioned problems in healthcare, education and making an income. Before the outbreak of Ebola there were also problems with water, safety and security and accessing food. Among those 6, at this time last year, what was 1. What was the biggest problem? ________________ 2. What was the 2
nd biggest problem? ______________
3. Which one was the 3rd
biggest problem?_____________
E2. Let’s talk about now, what are the top 3 problems for your county currently?
1. What is the biggest problem? ________________ 2. What is the 2
nd biggest problem? ______________
3. Which one is the 3rd
biggest?_____________
Maryland 173,866 15 12 14,489
Nimba 580,662 15 15 38,711
River Gee 83,565 10 11 7,597
Rivercess 89,470 10 11 8,134
Sinoe 129,296 15 12 10,775
Montserrado 1,427,229 30 32 44,601
Total 4,396,871 225 216 20,356
ACAPS - ENAP Liberia April 2015
65
The use of Borda counts outside the arena of public elections relies on idealizations. There are two that are of importance in our context, but one of them is more consequential for this Assessment:
The underlying voting model implies that the strength of preferences differs between any two adjacent options by the same amount. In addition, in the modified version, the preference for all non-elected candidates is equally low.
For example, if a key informant chooses water as the top priority (meaning: this is the sector with the most important unmet needs), food as the second, and safety/security as the third, the difference in the severity of unmet needs between water and food vs. that between food and safety/security are the same. This appears artificial, but ordinal judgments allow, for sufficiently large samples, underlying preferences to be inferred at interval level (for an almost arbitrarily selected sample application, see: Hatzinger and Dittrich 2012). The critical idealization in Borda is the equal difference at the level of the individual voter. In surveys, as a minimum precaution, one should demand that the interviewee understand, prior to answering, that she has exactly M choices. This was respected in the formulation of question E2 (see figure above), but not in E1. Second,
The use of the Borda count method is made on good faith that the preference order among items A, B, C, etc. is not affected by the inclusion, or not, of additional items U, V, W, etc.
For example, the relative priority of health and income generation should be the same, regardless of whether Questions E1 and E2 include "access to food" as an additional sector choice or not.
The challenge of "irrelevant alternatives" However, the Borda count is not robust to such changes; it does not secure "independence of clones" or "independence of irrelevant alternatives" (Wikipedia 2014a, 2014b). In plain English, access to food and income generation are in an end-and-means relationship. The addition of "food" to the sector choices therefore is liable to "steal votes" from the income generation option. Without the "food" choice, income generation might attract enough votes to outrank all others. It may do so with statistical significance in this sample survey of key informants. Thus, access to food is at best an irrelevant alternative (income generation implies access to food, barring the availability/accessibility aspect), or at worst a negative clone on income generation, in the logic of the Borda count. We will calculate this effect further below. Another modification that was not designed, but which naturally happened in the interviews arose from this: Not all key informants chose to designate three priorities. In this case, the scoring remained the same for the M < 3 exercised options as above, with all others scored 0. For example, among the 216 interviewees, 200 defined three current priority sectors, 15 exercised only two choices, and one chose only one sector. Without further investigation, we assume that the impact of the restricted choices was minor7. Despite its shortcomings, the Borda count is a satisfactory system to measure priorities, particularly on account of its simplicity. The mean Borda count is a quantity easily estimated for sample surveys; its uncertainty is visualized in confidence intervals; there are convenient tests for differences between means.
Select calculations
Priority sectors We are interested here primarily in comparisons of the mean Borda counts
for a given sector between the time before the crisis (spring 2014) and now (March - April 2014, when the data were collected)
for the six sectors together, in the same period of time (before vs. after the crisis). This chart succinctly summarizes the survey estimates. Dots designate the mean, spikes the confidence intervals. The red line at x = 1 is the expectation that, with three choices resulting in six votes, the mean over six sectors is = 1.
7 Statistical purists may expect that we also mention that the modified Borda score variables are nearly compositional. This means that, had all key informants exercised their three choices, for every informant 3 + 2 + 1 = 6 votes would be recorded. Since we do not pursue correlational analyses (such as factor analysis), we do not discuss this further. But it is commonsense that if some sectors achieve high mean Borda counts, others are bound to end up low - as in any election. Compositional analyses are demanding (Aitchison 1986; Thió-Henestrosa and Martín-Fernández 2005); potential users of the Assessment database should refrain from interpreting raw correlations among Borda score variables.
ACAPS - ENAP Liberia April 2015
66
Figure 4: Mean Borda counts for six sectors, before and after the crisis
The interpretation is given in the main body of the Assessment report. Numeric output is given in the section below.
"Access to food" as an irrelevant alternative We suspect that the addition of "access to food" in the list of sectors from which key informants were made to select three as priority sectors stole votes from "income generation" because of their close end/means-relationship (see above). We wish to estimate this effect. In a complete Borda count, with as many choices as sectors, this could be answered conclusively. All one would need to do is to create an alternative model, by excluding food, and bumping up by one all priority scores that had been lower than the score for food, then re-calculate and compare the models (the cited Wikipedia articles have didactic examples on such situations). For the modified count, this is not that simple. Some key informants may have been aware that they could state three priorities only (they were told so in Question E2), and some not, or not fully. Among the former, some may have emphasized the difficulty to survive by pressing both access to food and income generation. Others may have chosen only one of them, in order to have two choices left to state other priorities. Re-scoring in this situation does not produce valid comparisons. Instead, we attempt a comparison by combining food access and income generation. We argue that the combined priority can be approximated by the maximum of the two scores. If only one of them was chosen among the three priorities, then its score, by definition, is this maximum. If both were chosen, the larger score seems to be a fair measure of combined priority. The sum of the two would exaggerate the priority considerably since the other four sectors have no chance to be combined with any others in their scoring. The mean would understate it. We re-create a comparison plot like Figure 4, with "Access to Food" eliminated, and instead "Food/Inc" as a shorthand for the combined access to food / income generation priority.
ACAPS - ENAP Liberia April 2015
67
Figure 5: Mean Borda counts, before/after - food and income combined
Among the current priorities, Food/Inc dominates all others. Significance tests confirm that, but are a luxury, seeing the barely overlapping confidence intervals. There is therefore a strong suggestion, if not compelling evidence, that "access to food" acted as an irrelevant alternative. It stole votes from income generation. Its inclusion most likely distorted results. Rather than concluding that income generation currently is of similar priority to the people of Liberia as education and health, there is reason to believe that it outranks all other sectors that the Assessment investigated.
Statistical output We present minimally edited output from analyses run in the statistical application STATA (Stata Corporation 2011).
Priority sectors
Description
. * DESCRIBE BORDA SCORE VARIABLES: storage display value variable name type format label variable label ----------------------------------------------------------------------------------------- E1item1_1_ byte %10.0g Before: Education E1item2_1_ byte %10.0g Before: Food E1item3_1_ byte %10.0g Before: Health E1item4_1_ byte %10.0g Before: Income Generation E1item5_1_ byte %10.0g Before: Safety and Securit E1item6_1_ byte %10.0g Before: Water E2item1_2_ byte %10.0g Currently: Education E2item2_2_ byte %10.0g Currently: Food E2item3_2_ byte %10.0g Currently: Health E2item4_2_ byte %10.0g Currently: Income Generation E2item5_2_ byte %10.0g Currently: Safety and Securit E2item6_2_ byte %10.0g Currently: Water
Edu
catio
n
He
alth
Fo
od
/Inc
Safe
tyW
ate
r
Before
Now
Before
Now
Before
Now
Before
Now
Before
Now
0 .5 1 1.5 2Mean Borda count
Note: 216 key informants. Modified Borda counts, with three priority optionsfor each KI. Dots = mean counts; lines = 95%CI. Combined food/income, see text.
Model with food and income scores combined
Most important problems - before/now
ACAPS - ENAP Liberia April 2015
68
.
. * TABULATE RAW FREQUENCIES OF PRIORITIES: | Borda scores Variable | 0 1 2 3 | Total ----------------------+--------------------------------------------+---------- Before: Education | 71 52 53 40 | 216 Before: Food | 161 19 16 20 | 216 Before: Health | 70 52 63 31 | 216 Before: Income Genera | 110 33 26 47 | 216 Before: Safety and Se | 147 27 23 19 | 216 Before: Water | 107 18 33 58 | 216 Currently: Education | 72 43 44 57 | 216 Currently: Food | 154 26 15 21 | 216 Currently: Health | 68 48 65 35 | 216 Currently: Income Gen | 78 36 36 66 | 216 Currently: Safety and | 160 30 15 11 | 216 Currently: Water | 133 17 40 26 | 216 ----------------------+--------------------------------------------+---------- Total | 1,331 401 429 431 | 2,592 . * CALCULATE MEAN COUNTS with confidence intervals:
Before / after, all sectors
Before the crisis:
Survey: Mean estimation Number of strata = 15 Number of obs = 216 Number of PSUs = 216 Population size = 4396871 Design df = 201 -------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E1item1_1_ | 1.165241 .0895459 .9886712 1.341811 E1item2_1_ | .4439805 .06475 .3163041 .571657 E1item3_1_ | 1.384437 .0840772 1.218651 1.550224 E1item4_1_ | .9831686 .0948463 .7961471 1.17019 E1item5_1_ | .6017073 .075761 .4523189 .7510957 E1item6_1_ | 1.315709 .1037024 1.111225 1.520193 --------------------------------------------------------------
Currently:
Survey: Mean estimation Number of strata = 15 Number of obs = 216 Number of PSUs = 216 Population size = 4396871 Design df = 201 -------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E2item1_2_ | 1.330221 .0944998 1.143883 1.516559 E2item2_2_ | .4666225 .0636723 .3410712 .5921738 E2item3_2_ | 1.351736 .0862356 1.181694 1.521779 E2item4_2_ | 1.409692 .1011684 1.210204 1.609179 E2item5_2_ | .3911297 .0570752 .2785867 .5036727 E2item6_2_ | .9438223 .0883234 .769663 1.117982 --------------------------------------------------------------
Before / after, pairwise for each sector [same as above, but different arrangement for easier comparison]
ACAPS - ENAP Liberia April 2015
69
Education
-------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E1item1_1_ | 1.165241 .0895459 .9886712 1.341811 E2item1_2_ | 1.330221 .0944998 1.143883 1.516559 --------------------------------------------------------------
Access to food
-------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E1item2_1_ | .4439805 .06475 .3163041 .571657 E2item2_2_ | .4666225 .0636723 .3410712 .5921738 --------------------------------------------------------------
Health
-------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E1item3_1_ | 1.384437 .0840772 1.218651 1.550224 E2item3_2_ | 1.351736 .0862356 1.181694 1.521779 --------------------------------------------------------------
Income generation
-------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E1item4_1_ | .9831686 .0948463 .7961471 1.17019 E2item4_2_ | 1.409692 .1011684 1.210204 1.609179 --------------------------------------------------------------
Safety and security
-------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E1item5_1_ | .6017073 .075761 .4523189 .7510957 E2item5_2_ | .3911297 .0570752 .2785867 .5036727 --------------------------------------------------------------
Water
-------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] -------------+------------------------------------------------ E1item6_1_ | 1.315709 .1037024 1.111225 1.520193 E2item6_2_ | .9438223 .0883234 .769663 1.117982 --------------------------------------------------------------
Irrelevant alternatives: Combining food and income scores
Description storage display value variable name type format label variable label ----------------------------------------------------------------------------------------------- E1maxFoodIncome float %9.0g Before: Greater of food or income scores E2maxFoodIncome float %9.0g Currently: Greater of food or income scores
ACAPS - ENAP Liberia April 2015
70
Comparison with the previous, separate measures for food and income generation
----------------------------------------------------------------- | Linearized | Mean Std. Err. [95% Conf. Interval] ----------------+------------------------------------------------ Before: E1item2_1_ | .4439805 .06475 .3163041 .571657 E1item4_1_ | .9831686 .0948463 .7961471 1.17019 E1maxFoodIncome | 1.315863 .0972764 1.12405 1.507676 Currently: E2item2_2_ | .4666225 .0636723 .3410712 .5921738 E2item4_2_ | 1.409692 .1011684 1.210204 1.609179 E2maxFoodIncome | 1.730311 .0951778 1.542636 1.917987 ----------------------------------------------------------------- Tests show that the current Borda count of combined food/income is significantly higher than all the other sector counts (all p < 0.02, incl. vs. the separate income generation count). For the priorities before the crisis, it is significantly higher only compared to safety and security. Since Figure 5 makes these tests a foregone conclusion, the detailed tests are not reported, in order to save space.
STATA do-file The code is included in the interest of reproducibility. Interested researcher may approach ACAPS for a copy of the
data.
********************************************************************* * MEAN BORDA COUNTS * * Compare Before / currently; collect results in a graph * ********************************************************************* use "C:\..\150416_1529_Liberia_Borda_SurveyEst_LongVarLabels.dta", clear set more off * DESCRIBE BORDA SCORE VARIABLES: des E1item1_1_ - E1item6_1_ E2item1_2_ - E2item6_2_ * TABULATE RAW FREQUENCIES OF PRIORITIES: tabm E1item1_1_ - E1item6_1_ E2item1_2_ - E2item6_2_ /* tabm is a user-defined command, written by Nicholas J. Cox, University of Durham, U.K. */ * CALCULATE MEAN COUNTS: * Before the crisis: svy: mean E1item1_1_ - E1item6_1_ *Currently: svy: mean E2item1_2_ - E2item6_2_ * PLOTTING MULTIPLE MODELS IN ONE GRAPH (coefplot): svy: mean E1item1_1_ E2item1_2_ estimates store education svy: mean E1item2_1_ E2item2_2_ estimates store food svy: mean E1item3_1_ E2item3_2_ estimates store health svy: mean E1item4_1_ E2item4_2_ estimates store income svy: mean E1item5_1_ E2item5_2_ estimates store safety svy: mean E1item6_1_ E2item6_2_ estimates store water coefplot (education, aseq(Education) \ food, aseq(Food) \ health, aseq(Health) /// \ income, aseq("Income") \ safety, aseq("Safety") \ water, aseq(Water)), xline(1) xscale(range(0 2)) xlabel(0(0.5)2) /// xtitle(Mean Borda count) title(Most important problems) /// subtitle("Before the crisis vs. now") /// coeflabel( E1item1_1_ = "Before" E2item1_2_ = "Now" /// E1item2_1_ = "Before" E2item2_2_ = "Now" /// E1item3_1_ = "Before" E2item3_2_ = "Now" /// E1item4_1_ = "Before" E2item4_2_ = "Now" /// E1item5_1_ = "Before" E2item5_2_ = "Now" /// E1item6_1_ = "Before" E2item6_2_ = "Now") /// note("Note: 216 key informants. Modified Borda counts, with three priority options""for each KI. Dots = mean counts; lines = 95%CI. With six points distributed ""over the six sectors, the expected mean count is 1 (red line).")
ACAPS - ENAP Liberia April 2015
71
* The package "coefplot" was written by Ben Jann, University of Berne, Switzerland (Jann 2014). ********************************************************************* * TEST FOR "IRRELEVANT ALTERNATIVE" EFFECT * * of including "access to food" besides "income generation" * ********************************************************************* * CREATE COMBINED PRIORITY SCORE FOR "ACCESS TO FOOD" AND "INCOME GENERATION" capture egen E1maxFoodIncome = rowmax(E1item2_1_ E1item4_1_) label var E1maxFoodIncome "Before: Greater of food or income scores" capture egen E2maxFoodIncome = rowmax(E2item2_2_ E2item4_2_) label var E2maxFoodIncome "Currently: Greater of food or income scores" *Comparison with the previous, separate measures for food and income generation: svy: mean E1item2_1_ E1item4_1_ E1maxFoodIncome E2item2_2_ E2item4_2_ E2maxFoodIncome * TEST FOR SIGNIFICANT DIFFERENCES IN CURRENT PRIORITIES between the combined food/income vs. each of the others: * Before the crisis: svy: mean E1* foreach var of varlist E1item* { test E1maxFoodIncome -`var' = 0 } * Currently: svy: mean E2* foreach var of varlist E2item* { test E2maxFoodIncome -`var' = 0 } * PLOTTING MULTIPLE MODELS IN ONE GRAPH (replacing "food" and "income generation" with the combined food/income): svy: mean E1maxFoodIncome E2maxFoodIncome estimates store foodinc coefplot (education, aseq(Education) \ health, aseq(Health) /// \ foodinc, aseq("Food/Inc") \ safety, aseq("Safety") \ water, aseq(Water)), xline(1) xscale(range(0 2)) xlabel(0(0.5)2) /// xtitle(Mean Borda count) title("Most important problems - before/now") /// subtitle("Model with food and income scores combined") /// coeflabel( E1item1_1_ = "Before" E2item1_2_ = "Now" /// E1item3_1_ = "Before" E2item3_2_ = "Now" /// E1maxFoodIncome = "Before" E2maxFoodIncome = "Now" /// E1item5_1_ = "Before" E2item5_2_ = "Now" /// E1item6_1_ = "Before" E2item6_2_ = "Now") /// note("Note: 216 key informants. Modified Borda counts, with three priority options""for each KI. Dots = mean counts; lines = 95%CI. Combined food/income, see text.") set more on
References Aitchison, J. (1986). The statistical analysis of compositional data. Dordrecht Springer.
Benini, A. (2013). Severity and priority - Their measurement in rapid needs assessments Geneva, Assessment Capacity Project (ACAPS).
Borda, J. C. (1781). "Mémoire sur les élections au scrutin." Histoire de l’Académie royale des sciences 2: 85.
Hatzinger, R. and R. Dittrich (2012). "Prefmod: An R package for modeling preferences based on paired comparisons, rankings, or ratings." Journal of Statistical Software 48(10): 1-31.
Jann, B. (2014). "Plotting regression coefficients and other estimates." Stata Journal 14(4): 708-737.
Stata Corporation (2011). Stata Statistical Software, Release 12: Reference Manual. College Station, TX, StataCorp LP.
Swackhamer, D. L. and J. K. Hammit (2010). EPA’s Draft Expert Elicitation Task Force White Paper. Letter to the EPA Administrator, January 21, 2010 [EPA-SAB-10-003], Annex: Review of EPA’s Draft Expert Elicitation Task Force White Paper. U. E. P. A. Science Advisory Board. Washington DC.
Thió-Henestrosa, S. and J. Martín-Fernández (2005). "Dealing with compositional data: the freeware CoDaPack." Mathematical Geology 37(7): 773-793.
Wikipedia. (2011)."Borda count." Retrieved 28 March 2011, from http://en.wikipedia.org/wiki/Borda_count.
Wikipedia. (2014a)."Independence of clones criterion." Retrieved 25 April 2015, from http://en.wikipedia.org/wiki/Independence_of_clones_criterion.
Wikipedia. (2014b)."Independence of irrelevant alternatives." Retrieved 25 April 2015, from http://en.wikipedia.org/wiki/Independence_of_irrelevant_alternatives.